Sentiment Analysis on Twitter Social Media Regarding Covid-19 Vaccination with Naive Bayes Classifier (NBC) and Bidirectional Encoder Representations from Transformers (BERT)

Abstract. The Covid-19 vaccine is an important tool to stop the Covid-19 pandemic, however, there are pros and cons from the public regarding this Covid-19 vaccine. Purpose: These responses were conveyed by the public in many ways, one of which is through social media such as Twitter. Responses given by the public regarding the Covid-19 vaccination can be analyzed and categorized into responses with positive, neutral or negative sentiments. Methods: In this study, sentiment analysis was carried out regarding Covid-19 vaccination originating from Twitter using the Naïve Bayes Classifier (NBC) and Bidirectional Encoder Representations from Transformers (BERT) algorithms. The data used in this study is public tweet data regarding the Covid-19 vaccination with a total of 29,447 tweet data in English. Result: Sentiment analysis begins with data preprocessing on the dataset used for data normalization and data cleaning before classification. Then word vectorization was performed with TF-IDF and data classification was performed using the Naïve Bayes Classifier (NBC) and Bidirectional Encoder Representations from Transformers (BERT) algorithms. From the classification results, an accuracy value of 73% was obtained for the Naïve Bayes Classifier (NBC) algorithm and 83% for the Bidirectional Encoder Representations from Transformers (BERT) algorithm. Novelty: A direct comparison between classical models such as NBC and modern deep learning models such as BERT offers new insights into the advantages and disadvantages of both approaches in processing Twitter data. Additionally, this study proposes temporal sentiment analysis, which allows evaluating changes in public sentiment regarding vaccination over time. Another innovation is the implementation of a hybrid approach to data cleansing that combines traditional methods with the natural language processing capabilities of BERT, which more effectively addresses typical Twitter data issues such as slang and spelling errors. Finally, this research also expands sentiment classification to be multi-label, identifying more specific sentiment categories such as trust, fear, or doubt, which provides a deeper understanding of public opinion.

Sentiment analysis has come out as an important activity in natural language processing (NLP) applications whose data analysis is in high demand at present in the modern world. The BERT (Bidirectional Encoder Representations from Transformers) algorithm has proved to be extremely efficient when it comes to sentiment analysis tasks, and its potential is far exceeding that of conventional algorithms, unlocking their potential however would require fine tuning of their hyperparameters. It is quite a feat to optimise the BERT’s various hyperparameters due to the complicated interaction between them (e.g. the learning rate, batch size, dropout rate, attention heads). In this paper, the Salp Swarm Algorithm (SSA) is used as a bio-inspired metaheuristic optimization technique to optimize the fine-tuning process. Through SSA’s exceptionally efficient search capabilities in modelling multidimensional search space, BERT hyperparameters are optimized systematically to the sentiment classification tasks. A benchmark dataset for sentiment analysis (Sentiment140) is used to evaluate the proposed model. The novelty of the presented model is the fact that it dynamically adjusts its search behaviour in response to performance signals, thus it identifies better-performing parameter sets than conventional methods, leading to successful exploitation of the BERT algorithm that has produced high performing configurations. Extensive evaluations against 3 state-of-the-art search algorithms, namely manual tuning, grid search, and random search are conducted on the Sentiment140 benchmark dataset, demonstrating the superiority of the proposed SSA BERT optimization technique over state-of-the-art methods. The SSA-BERT model achieved a maximum accuracy of 96.4 percent, which is far better than manual tuning, grid search, and random search (65.0 percent, 69.5 percent and 72.0 percent respectively). It also performed better than other existing BERT models used in related literature, which showed accuracy levels between 46.4 and 75.7 percent in accordance with different benchmarks Sentiment analysis has come out as an important activity in natural language processing (NLP) applications whose data analysis is in high demand at present in the modern world. The BERT (Bidirectional Encoder Representations from Transformers) algorithm has proved to be extremely efficient when it comes to sentiment analysis tasks, and its potential is far exceeding that of conventional algorithms, unlocking their potential however would require fine tuning of their hyperparameters. It is quite a feat to optimise the BERT’s various hyperparameters due to the complicated interaction between them (e.g. the learning rate, batch size, dropout rate, attention heads). In this paper, the Salp Swarm Algorithm (SSA) is used as a bio-inspired metaheuristic optimization technique to optimize the fine-tuning process. Through SSA’s exceptionally efficient search capabilities in modelling multidimensional search space, BERT hyperparameters are optimized systematically to the sentiment classification tasks. A benchmark dataset for sentiment analysis (Sentiment140) is used to evaluate the proposed model. The novelty of the presented model is the fact that it dynamically adjusts its search behaviour in response to performance signals, thus it identifies better-performing parameter sets than conventional methods, leading to successful exploitation of the BERT algorithm that has produced high performing configurations. Extensive evaluations against 3 state-of-the-art search algorithms, namely manual tuning, grid search, and random search are conducted on the Sentiment140 benchmark dataset, demonstrating the superiority of the proposed SSA BERT optimization technique over state-of-the-art methods. The SSA-BERT model achieved a maximum accuracy of 96.4 percent, which is far better than manual tuning, grid search, and random search (65.0 percent, 69.5 percent and 72.0 percent respectively). It also performed better than other existing BERT models used in related literature, which showed accuracy levels between 46.4 and 75.7 percent in accordance with different benchmarks.

This study analyzes the comparison of two algorithms, namely Naive Bayes and Bidirectional Encoder Representations From Transformers (BERT), for the evaluation of the performance of education personnel at SMK MULTI KARYA This study uses manual calculation methods and the Python application. The results showed that the Naive Bayes algorithm gave very consistent results with accuracy, precision, and recall values of 76.67% both in manual calculations and with Pyton. This indicates that the Naive Bayes algorithm is effective in grouping data on the performance of education personnel. Meanwhile, the Bidirectional Encoder Representations From Transformers (BERT) algorithm shows mixed results, while with Python it reaches 12.00%. There are significant differences in recall values and precision between these two calculation methods. Nevertheless, the performance category "Good Performance Staff" remains the most dominant. The difference in results between manual and python calculations is that Naive bayes is a more stable and consistent method across different platforms, whereas Bidirectional Encoder Representations From Transformers (BERT) shows flexibility but with smaller variation in results. Therefore, in the context of education performance evaluation, NAive bayes are more reliable to produce consistent performance categories, while Bidirectional Encoder Representations From Transformers(BERT) can be an alternative with a fairly high level of accuracy but require further consideration in the interpretation of the results..

Text classification is a fundamental task in Natural Language Processing (NLP) with a wide range of applications such as sentiment analysis, document classification and content recommendation. Traditional approaches like Naive Bayes (NB), Support Vector Machine (SVM) and Random Forest (RF) relied on feature engineering but lacked contextual understanding. Deep learning came into the picture for text classification with transformer models such as Bidirectional Encoder Representations from Transformers (BERT), which could understand contextual words bidirectionally. In this article, we utilize a pre-trained BERT model fine-tuned on the Reuters-21578 dataset to classify news articles. We aim to measure the performance of transfer learning against common machine learning models and non-fine-tuned BERT. The fine-tuned model achieves 91.77% accuracy, which significantly outperforms the non-fine-tuned BERT, and performs better than classical classifiers such as NB, SVM and RF. The results show that fine-tuning allows BERT to contextualize domain-specific intricacies, resulting in improved classification performance. We also address the computational trade-offs associated with transformer models, highlighting the need for optimal methods for deployment. Thus, this study further enables the use of fine-tuned BERT in automatic news classification and is of significant value for information retrieval and content personalization.

Since the spread of the coronavirus flu in 2019 (hereafter referred to as COVID-19), millions of people worldwide have been affected by the pandemic, which has significantly impacted our habits in various ways. In order to eradicate the disease, a great help came from unprecedentedly fast vaccines development along with strict preventive measures adoption like lockdown. Thus, world wide provisioning of vaccines was crucial in order to achieve the maximum immunization of population. However, the fast development of vaccines, driven by the urge of limiting the pandemic caused skeptical reactions by a vast amount of population. More specifically, the people’s hesitancy in getting vaccinated was an additional obstacle in fighting COVID-19. To ameliorate this scenario, it is important to understand people’s sentiments about vaccines in order to take proper actions to better inform the population. As a matter of fact, people continuously update their feelings and sentiments on social media, thus a proper analysis of those opinions is an important challenge for providing proper information to avoid misinformation. More in detail, sentiment analysis (Wankhade et al. in Artif Intell Rev 55(7):5731–5780, 2022. https://doi.org/10.1007/s10462-022-10144-1) is a powerful technique in natural language processing that enables the identification and classification of people feelings (mainly) in text data. It involves the use of machine learning algorithms and other computational techniques to analyze large volumes of text and determine whether they express positive, negative or neutral sentiment. Sentiment analysis is widely used in industries such as marketing, customer service, and healthcare, among others, to gain actionable insights from customer feedback, social media posts, and other forms of unstructured textual data. In this paper, Sentiment Analysis will be used to elaborate on people reaction to COVID-19 vaccines in order to provide useful insights to improve the correct understanding of their correct usage and possible advantages. In this paper, a framework that leverages artificial intelligence (AI) methods is proposed for classifying tweets based on their polarity values. We analyzed Twitter data related to COVID-19 vaccines after the most appropriate pre-processing on them. More specifically, we identified the word-cloud of negative, positive, and neutral words using an artificial intelligence tool to determine the sentiment of tweets. After this pre-processing step, we performed classification using the BERT + NBSVM model to classify people’s sentiments about vaccines. The reason for choosing to combine bidirectional encoder representations from transformers (BERT) and Naive Bayes and support vector machine (NBSVM ) can be understood by considering the limitation of BERT-based approaches, which only leverage encoder layers, resulting in lower performance on short texts like the ones used in our analysis. Such a limitation can be ameliorated by using Naive Bayes and Support Vector Machine approaches that are able to achieve higher performance in short text sentiment analysis. Thus, we took advantage of both BERT features and NBSVM features to define a flexible framework for our sentiment analysis goal related to vaccine sentiment identification. Moreover, we enrich our results with spatial analysis of the data by using geo-coding, visualization, and spatial correlation analysis to suggest the most suitable vaccination centers to users based on the sentiment analysis outcomes. In principle, we do not need to implement a distributed architecture to run our experiments as the available public data are not massive. However, we discuss a high-performance architecture that will be used if the collected data scales up dramatically. We compared our approach with the state-of-art methods by comparing most widely used metrics like Accuracy, Precision, Recall and F-measure. The proposed BERT + NBSVM outperformed alternative models by achieving 73% accuracy, 71% precision, 88% recall and 73% F-measure for classification of positive sentiments while 73% accuracy, 71% precision, 74% recall and 73% F-measure for classification of negative sentiments respectively. These promising results will be properly discussed in next sections. The use of artificial intelligence methods and social media analysis can lead to a better understanding of people’s reactions and opinions about any trending topic. However, in the case of health-related topics like COVID-19 vaccines, proper sentiment identification could be crucial for implementing public health policies. More in detail, the availability of useful findings on user opinions about vaccines can help policymakers design proper strategies and implement ad-hoc vaccination protocols according to people’s feelings, in order to provide better public service. To this end, we leveraged geospatial information to support effective recommendations for vaccination centers.

First developed in 2018 by Google researchers, Bidirectional Encoder Representations from Transformers (BERT) represents a breakthrough in natural language processing (NLP). BERT achieved state-of-the-art results across a range of NLP tasks while using a single transformer-based neural network architecture. This work reviews BERT's technical approach, performance when published, and significant research impact since release. We provide background on BERT's foundations like transformer encoders and transfer learning from universal language models. Core technical innovations include deeply bidirectional conditioning and a masked language modeling objective during BERT's unsupervised pretraining phase. For evaluation, BERT was fine-tuned and tested on eleven NLP tasks ranging from question answering to sentiment analysis via the GLUE benchmark, achieving new state-of-the-art results. Additionally, this work analyzes BERT's immense research influence as an accessible technique surpassing specialized models. BERT catalyzed adoption of pretraining and transfer learning for NLP. Quantitatively, over 10,000 papers have extended BERT and it is integrated widely across industry applications. Future directions based on BERT scale towards billions of parameters and multilingual representations. In summary, this work reviews the method, performance, impact and future outlook for BERT as a foundational NLP technique. We provide background on BERT's foundations like transformer encoders and transfer learning from universal language models. Core technical innovations include deeply bidirectional conditioning and a masked language modeling objective during BERT's unsupervised pretraining phase. For evaluation, BERT was fine-tuned and tested on eleven NLP tasks ranging from question answering to sentiment analysis via the GLUE benchmark, achieving new state-of-the-art results. Additionally, this work analyzes BERT's immense research influence as an accessible technique surpassing specialized models. BERT catalyzed adoption of pretraining and transfer learning for NLP. Quantitatively, over 10,000 papers have extended BERT and it is integrated widely across industry applications. Future directions based on BERT scale towards billions of parameters and multilingual representations. In summary, this work reviews the method, performance, impact and future outlook for BERT as a foundational NLP technique.

In the ever-changing world of financial markets, understanding investor behavior and making informed decisions relies heavily on sentiment analysis. This study delves into the integration of traditional techniques, such as the Loughran- McDonald dictionary, with advanced natural language processing (NLP) methods utilizing BERT (Bidirectional Encoder Representations from Transformers). The goal is to enhance the accuracy and depth of sentiment analysis in financial reports.To begin, we employ the specialized Loughran-McDonald dictionary designed for financial sentiment analysis. This lexicon includes domainspecific word lists for positive and negative sentiments, forming a solid foundation for sentiment scoring. Expanding on this foundation, we incorporate BERT, an advanced transformerbased NLP model. BERT’s contextual understanding of language and ability to capture intricate semantic relationships within financial texts aim to overcome the limitations of rule-based sentiment analysis. The methodology involves preprocessing financial reports, integrating Loughran-McDonald sentiment scores, and fine-tuning BERT for financial sentiment classification. This hybrid approach leverages both the domain expertise encoded in the dictionary and BERT’s contextual comprehension of financial jargon and nuances. We validate and evaluate our implementation using a diverse dataset comprising quarterly earnings releases, annual reports, and other relevant disclosures. Performance metrics such as precision, recall, and F1 score are analyzed to assess the effectiveness of our hybrid approach compared to individual methods. The findings have significant implications for financial analysts, investors, and policymakers by providing a more nuanced understanding of sentiment in financial reports. Our hybrid approach aims to offer improved accuracy in capturing sentiment polarity while facilitating more informed decision-making in today’s complex and dynamic realm of financial markets.

As social media and online social networking sites have emerged as essential channels for everyday social interactions, researchers are increasingly turning to conversational social media as a valuable alternative source of insights into public sentiment and behavior during times of crisis. In a similar vein, our study investigates the utility of using conversational social media data for analyzing public sentiment related to the COVID-19 pandemic. Our research question is, to what extent can machine learning models based on conversational social media data help gauge public sentiment during a pandemic? To evaluate the performance of various machine learning algorithms in sentiment analysis tasks, we used a Twitter dataset and compared their effectiveness using eight different models (i.e., Decision Tree (DT), Random Forest (RF), Naïve Bayes (NB), Extreme Gradient Boosting (XGBoost), Logistic Regression, Support Vector Machine (SVM), Long Short-Term Memory (LSTM), and Bidirectional Encoder Representations from Transformers (BERT)). The results show that among the eight models, the top three models are BERT, LSTM, and SVM, achieving overall accuracies of 87.44%, 81.94%, and 80.02%, respectively. Employing sentiment analysis models on pandemic-related social media posts allows us to gain insights into of human emotions and sentiments during public health crises like a pandemic.

Sentiments are the key factors that lead to influence our behavior. Sentiment analysis is a technique that analyzes people's behaviors, attitudes, and emotions toward a service, product, topic, or event. Since 2020, no country has remained untouched by COVID-19, and the governing bodies of most countries have been applying several anti-pandemic countermeasures to combat it. In this regard, it becomes tremendously important to analyze people's sentiments when tackling infectious diseases similar to COVID-19. The countermeasures taken by any country to control the pandemic leave a direct and crucial impact on each sector of public life, and every individual reacts to them differently. It is necessary to consider these reactions to implement appropriate messaging and decisive policies. Pakistan has done enough to control this virus's spread like every other country. This research aims to perform a sentimental analysis on the famous microblogging social platform, Twitter, to get insights into public sentiments and the attitudes displayed towards the precautionary steps taken by the Government of Pakistan in the years 2020 and 2021. These steps or countermeasures include the closure of educational institutes, suspension of flight operations, lockdown of business activities, enforcement of several standard operating procedures (SOPs), and the commencement of the vaccination program. We implemented four approaches for the analysis, including the Valence Aware Dictionary and sEntiment Reasoner (VADER), TextBlob, Flair, and Bidirectional Encoder Representations from Transformers (BERT). The first two techniques are lexicon-based. Flair is a pre-trained embedding-based approach, whereas BERT is a transformer-based model. BERT was fine-tuned and trained on a labeled dataset, achieving a validation accuracy of 92%. We observed that the polarity score kept varying from month to month in both years for all countermeasures. This score was analyzed with real-time events occurring in the country, which helped understand the public's sentiment and led to the possible formation of a notable conclusion. All implemented approaches showed independent performances. However, we noticed from the classification results of both TextBlob and the fine-tuned BERT model that neutral sentiment was dominant in the data, followed by positive sentiment.

General background Israel-Palestine conflict has drawn significant global attention, particularly in how it is perceived and discussed on social media platforms. Specific background understanding public sentiment surrounding such geopolitical issues is crucial for media monitoring, diplomatic efforts, and reputation management. Knowledge gap previous sentiment analysis studies often lack the ability to accurately handle multilingual and context-rich datasets, especially in analyzing neutral sentiments, which are commonly overlooked. This study aims to apply the Bidirectional Encoder Representations from Transformers (BERT) model to analyze public sentiment towards the Israel-Palestine conflict on the X platform, focusing on Indonesian users. Results Using BERT, the model achieved 93% accuracy, with a precision of 0.95, recall of 0.93, and F1-score of 0.94. The model performed well in predicting positive and negative sentiments but showed room for improvement in handling neutral sentiment. Novelty this study introduces the implementation of the BERT Transformer model for the multilingual and context-sensitive sentiment analysis of tweets, specifically addressing a high-stakes geopolitical conflict. Implications the findings demonstrate the potential for using advanced natural language processing techniques like BERT for monitoring public opinion, brand management, and detecting societal tensions on social media, offering valuable insights for stakeholders involved in conflict resolution and diplomatic strategies. Highlights: Achieved 93% accuracy in sentiment analysis using BERT on X platform. Identified strengths in predicting positive/negative sentiments, with challenges in neutral sentiment. Demonstrated BERT’s effectiveness in handling complex geopolitical social media data. Keywords: Sentiment Analysis, BERT Transformer, Israel-Palestine Conflict, Social Media, NLP

Objectives : This work explores the application of two advanced state-of-the-art models, BERT (Bidirectional Encoder Representations from Transformers) and USE (Universal Sentence Encoder), to automate the grading of short answers. Methods: This work investigates the use of BERT and USE models for automatically grading short answers. The research utilizes HP:SAS dataset containing manually graded responses by two human evaluators. The student responses as well as model answers responses of question 1 and question set 6 are then processed using the BERT and USE models, with scores generated based on cosine similarity measures between student answers and predefined model answers. Findings: The work demonstrates that BERT and USE embeddings can effectively capture contextual and semantic similarity, their performance is heavily dependent on the function which generates the score. Our finding reveal that a non-linear mapping function mimics the human grading more than a linear mapping function. Such a function enhances accuracy (0.67) and reduces the error (0.617) by computing Pearson correlation coefficient and RMSE respectively. Notably, longer responses achieved higher Pearson correlations (0.67) than shorted answers (0.59). The results bring out usability and choice aspects of BERT and USE in relation to ASAG, contributing to the understanding of their application across various answers. We conclude with a weighted ensemble method combining BERT and USE with subject- specific strictness parameter (k) provides a robust framework for automated assessment. Novelty: Evaluates and compares two deep learning models for automatic short answer grading, a scarcely explored area. A novel contribution is the granular analysis across different scoring ranges across two question sets of the dataset. The novelty of this work lies in the transition from linear scoring to non-linear mapping framework. This approach introduces a tunable sigmoid- based ensemble that would replicate human assessment. Finally, a comparison with existing studies demonstrates very limited research. Keywords: Bidirectional Encoder Representations from Transformers (BERT), Universal Sentence Encoder (USE), Transformer, word embedding, Non-linear mapping, deep learning, short answer grading

Recent research has focused on opinion mining from public sentiments using natural language processing (NLP) and machine learning (ML) techniques. Transformer-based models, such as bidirectional encoder representations from transformers (BERT), excel in extracting semantic information but are resourceintensive. Google’s new research, mixing tokens with fourier transform, also known as FNet, replaced BERT’s attention mechanism with a non-parameterized fourier transform, aiming to reduce training time without compromising performance. This study fine-tuned the FNet model with a publicly available Kaggle hotel review dataset and investigated the performance of this dataset in both FNet and BERT architectures along with conventional machine learning models such as long short-term memory (LSTM) and support vector machine (SVM). Results revealed that FNet significantly reduces the training time by almost 20% and memory utilization by nearly 60% compared to BERT. The highest test accuracy observed in this experiment by FNet was 80.27% which is nearly 97.85% of BERT’s performance with identical parameters.

With the steady growth of social media and online platforms, sentiment analysis has become a critical task to understand public opinion, customer feedback, and social trends. This study investigates multi-modal sentiment analysis by exploring state-of-the-art models such as Bidirectional encoder representations from transformers (BERT), Bootstrapping Language-Image Pretraining (BLIP), Generative Image-to-text Transformer (GIT), and Contrastive Language-Image Pretraining (CLIP) for sentiment classification tasks using both image and text data. The effectiveness of these models on sentiment analysis tasks is evaluated under different configurations such as text-only BERT, image-to-text augmented BERT, and CLIP-based classification. The results show that while BERT achieves 76% accuracy on text-only sentiment analysis, combining text with image-generated descriptions does not significantly improve performance, with accuracy remaining around 74%. On the other hand, CLIP achieves a moderate 62% accuracy using image-text embeddings. While CLIP performs well on multi-modal mapping, it demonstrates challenges in deep semantic understanding compared to BERTs 0.76 F1 score on the text-only task. These findings highlight the challenges of effectively merging different modalities and point out future directions for improving sentiment analysis in multi-modal settings, enhancing the ability of models to fully understand the semantic content in both images and text.

This study examines how sentiment analysis of environmental, social, and governance (ESG) news affects the financial performance of companies in innovative sectors such as mobility, technology, and renewable energy. Using approximately 9828 general ESG articles from Google News and approximately 140,000 company-specific ESG articles, we performed term frequency-inverse document frequency (TF-IDF) analysis to identify key ESG-related terms and visualize their materiality across industries. We then applied models such as bidirectional encoder representations from transformers (BERT), the robustly optimized BERT pretraining approach (RoBERTa), and big bidirectional encoder representations from transformers (BigBird) for multiclass sentiment analysis, and distilled BERT (DistilBERT), a lite BERT (ALBERT), tiny BERT (TinyBERT), and efficiently learning an encoder that classifies token replacements accurately (ELECTRA) for positive and negative sentiment identification. Sentiment analysis results were correlated with profitability, cash flow, and stability indicators over a three-year period (2019–2021). ESG ratings from Morgan Stanley Capital International (MSCI), a prominent provider that evaluates companies’ sustainability practices, further enriched our analysis. The results suggest that sentiment impacts financial performance differently across industries; for example, positive sentiment correlates with financial success in mobility and renewable energy, while consumer goods often show positive sentiment even with low environmental ESG scores. The study highlights the need for industry-specific ESG strategies, especially in dynamic sectors, and suggests future research directions to improve the accuracy of ESG sentiment analysis.

BackgroundGiven the increasing number of people suffering from tinnitus, the accurate categorization of patients with actionable reports is attractive in assisting clinical decision making. However, this process requires experienced physicians and significant human labor. Natural language processing (NLP) has shown great potential in big data analytics of medical texts; yet, its application to domain-specific analysis of radiology reports is limited.ObjectiveThe aim of this study is to propose a novel approach in classifying actionable radiology reports of tinnitus patients using bidirectional encoder representations from transformer BERT-based models and evaluate the benefits of in domain pre-training (IDPT) along with a sequence adaptation strategy.MethodsA total of 5864 temporal bone computed tomography(CT) reports are labeled by two experienced radiologists as follows: (1) normal findings without notable lesions; (2) notable lesions but uncorrelated to tinnitus; and (3) at least one lesion considered as potential cause of tinnitus. We then constructed a framework consisting of deep learning (DL) neural networks and self-supervised BERT models. A tinnitus domain-specific corpus is used to pre-train the BERT model to further improve its embedding weights. In addition, we conducted an experiment to evaluate multiple groups of max sequence length settings in BERT to reduce the excessive quantity of calculations. After a comprehensive comparison of all metrics, we determined the most promising approach through the performance comparison of F1-scores and AUC values.ResultsIn the first experiment, the BERT finetune model achieved a more promising result (AUC-0.868, F1-0.760) compared with that of the Word2Vec-based models(AUC-0.767, F1-0.733) on validation data. In the second experiment, the BERT in-domain pre-training model (AUC-0.948, F1-0.841) performed significantly better than the BERT based model(AUC-0.868, F1-0.760). Additionally, in the variants of BERT fine-tuning models, Mengzi achieved the highest AUC of 0.878 (F1-0.764). Finally, we found that the BERT max-sequence-length of 128 tokens achieved an AUC of 0.866 (F1-0.736), which is almost equal to the BERT max-sequence-length of 512 tokens (AUC-0.868,F1-0.760).ConclusionIn conclusion, we developed a reliable BERT-based framework for tinnitus diagnosis from Chinese radiology reports, along with a sequence adaptation strategy to reduce computational resources while maintaining accuracy. The findings could provide a reference for NLP development in Chinese radiology reports.