Pith. sign in

REVIEW 4 major objections 4 minor 52 references

Visualizing Public Opinion on X: A Real-Time Sentiment Dashboard Using VADER and DistilBERT

T0 review · 4 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that a weighted ensemble of VADER and DistilBERT beats both individual models on filtered English tweet sentiment, reaching 87.6% accuracy and F1 0.841, and that this hybrid index ranks Amazon far above Microsoft.

desk verdict The ensemble equation inverts polarity for confident negative tweets, so the reported 87.6% accuracy cannot be produced by the method as written; a well-structured student project, but not a publishable result. read the letter →

arxiv 2504.15448 v2 pith:NNOI5AHX submitted 2025-04-21 econ.GN cs.CLq-fin.EC

Yanampally Abhiram Reddy , Siddhi Agarwal , Vikram Parashar , Arshiya Arora This is my paper · ORCID
classification econ.GNcs.CLq-fin.EC
keywords SentimentanalysisVADERDistilBERTEnsembleTwitter/XCorporatereputationmonitoringReal-timeNLPIndex
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that a simple weighted combination of a rule-based sentiment lexicon (VADER) and a fine-tuned transformer (DistilBERT) classifies sentiment in filtered English posts on X more accurately than either model by itself, reaching 87.6% accuracy and an F1 score of 0.841. The aim is a real-time corporate reputation dashboard: each tweet is scored, the scores are averaged into a 0-100 Corporate Sentiment Index per company, and the index ranks public perception of twenty major corporations. If the hybrid claim holds, it would mean that a cheap, interpretable ensemble can deliver contextual accuracy close to a full BERT model at roughly a third of its processing time, making continuous sentiment monitoring practical. The paper applies this pipeline to 10,000 company-related tweets and reports sharp sentiment gaps, with Amazon at 81.2 and Microsoft at 21.7.

What carries the argument

The load-bearing object is the weighted ensemble equation, $S = \alpha S_{\text{VADER}} + (1-\alpha) S_{\text{DistilBERT}}$, with $\alpha=0.4$ found by grid search on the validation set. $S_{\text{VADER}}$ is the normalized compound score (scaled from -1..1 to 0..1), and $S_{\text{DistilBERT}}$ is the probability of the highest sentiment class. This equation carries the argument because it converts two very different outputs into one number that can be thresholded into positive, neutral, or negative labels and averaged into the Corporate Sentiment Index. The paper also relies on a preprocessing pipeline that removes URLs, expands emojis and slang, and lemmatizes for VADER while skipping such normalization for DistilBERT; this asymmetry is what lets the two models keep their distinct strengths before the weights recombine them.

What would settle it

Label a new batch of English tweets about the same companies from a later date, run the pipeline with alpha fixed at 0.4 and the already fine-tuned DistilBERT unchanged, and compare hybrid accuracy with VADER-only and DistilBERT-only on that batch; if the hybrid does not beat both, the ensemble claim fails. A cheaper check is to re-split the existing 2,000 labeled tweets into tuning and test halves and re-measure: accuracy falling below 0.843 would indicate the reported gain is in-sample fit.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that the hybrid score $S = \alpha S_{\text{VADER}} + (1-\alpha) S_{\text{DistilBERT}}$ with $\alpha = 0.4$ outperforms both single-model baselines, with accuracy 0.876 and F1 0.841, and does so at 47.5 ms per tweet, between VADER's 12.3 ms and DistilBERT's 78.6 ms. The paper traces the gain to complementary error patterns: VADER handles emojis, abbreviations, and sentiment-bearing lexical items, while DistilBERT captures context, and the weighted average compensates for each model's weaknesses. It then derives the Corporate Sentiment Index, $\text{CSI} = 100 \cdot \frac{1}{n}\sum_{i=1}^n S_i$, from the ensemble scores, and uses it to divide twenty companies into sentiment tiers: Amazon 81.2 and Samsung 45.8 as strongly positive; Microsoft 21.7 and Walmart 21.9 as poor. The discovery is therefore both methodological (better classification through fusion) and applied (public sentiment toward corporations is highly heterogeneous and not aligned with market capitalization).

Load-bearing premise

The load-bearing assumption is that the 87.6% accuracy and 0.841 F1 were measured on tweets held out from the tuning that set alpha to 0.4 and from the fine-tuning of DistilBERT; the paper never documents such a split.

Editorial extensions

If this is right

  • If the central accuracy claim holds, brand teams could run continuous monitoring at roughly 20 tweets per second per machine, fast enough for a live reputation dashboard.
  • The reported tiers imply that reputation management is targetable: Microsoft's low index is tied to privacy and software reliability complaints, while Amazon's high index coexists with negative clusters on labor practices and environmental impact.
  • The observed gap between sentiment and market value for firms like Microsoft implies investors should treat public perception as information independent of price, not as a mirror of financial performance.
  • At 47.5 ms per tweet, the hybrid sits between VADER and DistilBERT in speed while beating both in reported accuracy, which is the exact efficiency-accuracy trade-off the paper argues real-time systems need.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the fusion's 3.3-point gain over DistilBERT-only assumes VADER and DistilBERT make different, weakly correlated errors on this tweet filter; on sarcastic, code-mixed, or longer text the error overlap may grow and erase the gain.
  • Editorial inference: the CSI tiers are point means without confidence intervals, so pairs like Samsung 45.8 vs UnitedHealth 44.3 are likely within sampling noise; bootstrapping per-company tweet samples would check whether the four tiers survive.
  • Editorial inference: the paper's sentiment-valuation disconnect is a testable trading hypothesis; comparing weekly CSI changes with subsequent abnormal stock returns would show whether sentiment drift carries information price does not.
  • Editorial inference: because the threshold mapping (positive at 0.6, negative at 0.4) is applied to a blend of a normalized lexicon score and a class probability, the two inputs have different scales; calibrating the DistilBERT probability before fusing could shift the optimal alpha and thresholds.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper proposes a real-time sentiment-analysis system for X (formerly Twitter) that combines a rule-based model (VADER) with a transformer model (DistilBERT) in a weighted ensemble. The authors describe a web-scraping pipeline, a preprocessing chain, an ensemble classification rule (Eq. 1), and a Corporate Sentiment Index (CSI, Eq. 2). They report that the hybrid approach achieves 87.6% accuracy and F1 0.841, outperforming VADER-only, DistilBERT-only, BERT-base, and TextBlob baselines (Table I). The framework is then applied to 10,000 tweets about 20 corporations (500 per company), with a manually labeled 2,000-tweet subset, to produce corporate sentiment rankings (e.g., Amazon 81.2, Microsoft 21.7) and a four-tier sentiment stratification.

Significance. If the central performance claims were valid, the paper would offer a useful engineering contribution: a real-time, interpretable hybrid sentiment pipeline with a head-to-head comparison against stronger baselines and a practical corporate-monitoring case study. The explicit comparison of accuracy, F1, and processing time, including a full BERT baseline, is a strength, as is the attempt to combine model types for speed and contextual accuracy. However, the reported results are undermined by a definitional flaw in the ensemble equation that inverts the polarity of confident predictions, and by the absence of any described held-out test set. Because both the headline accuracy and the corporate sentiment rankings depend directly on these two issues, the central claims of the paper are not supported as stated.

major comments (4)
  1. [Sec. III-C3, Eq. (1)] The definition of S_DistilBERT as 'the probability of the highest sentiment class' discards the class label, so a high-confidence negative prediction from DistilBERT contributes a high value to the weighted sum. With α=0.4, a tweet with neutral VADER (S_VADER=0.5) and a DistilBERT negative probability of 0.9 gives S = 0.4·0.5 + 0.6·0.9 = 0.74, which is classified as positive under the S≥0.6 threshold. This is a polarity inversion for exactly the confident predictions the ensemble is supposed to exploit. The per-class accuracies in Table I (negative 87.3%, positive 91.2%) and the headline 87.6% accuracy are therefore incompatible with the method as written, even on a perfectly held-out split.
  2. [Sec. IV-A1 and Sec. III-C3] No held-out test split is described. The 2,000-tweet labeled subset is used both to fine-tune DistilBERT (validation accuracy 84.3%) and to grid-search the ensemble weight α ('optimal value ... determined through grid search on our validation set'), and the same labeled pool appears to be the basis for the accuracy and F1 reported in Table I. The reported 87.6%/0.841 figures are thus in-sample measures of the fitted pipeline, not estimates of generalization. The paper needs a clear train/validation/test partition with evaluation on an unseen test set before any claim of outperforming the baselines can be credited.
  3. [Sec. IV-C, Fig. 2] The sentiment tier stratification contradicts the reported CSI values. The 'Excellent Sentiment (CSI>40)' tier lists Amazon (81.2) and Samsung (45.8), but the 'Good Sentiment (CSI 35–39.9)' tier lists UnitedHealth (44.3) and NVIDIA (42.9), both of which exceed 40 and should be in the Excellent tier. The tier boundaries and the assigned companies are internally inconsistent, so the claimed 'natural clustering into four sentiment tiers' is not supported by the numbers presented.
  4. [Sec. IV-C and Eq. (2)] The Corporate Sentiment Index is presented as evidence of 'significant disparities' across companies (e.g., Amazon 81.2 vs. Microsoft 21.7), but no confidence intervals, standard errors, or statistical tests are provided for these company-level scores. Since each company's CSI is an average over only 500 tweets, the ranking could be dominated by noise or by the polarity-inverted ensemble discussed above; the case-study conclusions are not statistically supported as reported.
minor comments (4)
  1. [Sec. III-C3, Eq. (1)] The notation S_VADER is described as a normalized compound score scaled from -1..1 to 0..1, but the text does not state the mapping explicitly. A concrete formula or reference to the normalization transform would improve reproducibility.
  2. [Sec. V-D and Sec. VI-B] The limitations sections are candid about language constraints and sentiment simplification, but they do not mention the two most serious threats to validity identified above: the ensemble formulation and the lack of a test split.
  3. [Supplementary Material] The supplementary material link is given as 'IEEEtran', which appears to be a placeholder rather than an actual URL or repository reference.
  4. [References] Several references are incomplete or point to non-archival secondary sources (e.g., [15], [20], [45], [47]); full citation details and archival versions would help readers verify the stated related work.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline accuracy is measured on the validation set used to tune the ensemble weight and fine-tune DistilBERT, so the reported 87.6% is a fitted in-sample number rather than an independent prediction.

  1. fitted input called prediction [Sec. III-C3 (Eq. 1), Sec. III-C2, Sec. IV-A1, Sec. IV-B (Table I)]
    "The optimal value of α was determined through grid search on our validation set, with the final system using α = 0.4, giving slightly more weight to the contextual understanding of DistilBERT... The model achieved 84.3% accuracy on our validation set. ... Our hybrid approach achieved the highest accuracy (87.6%) and F1-score (0.841), outperforming both individual models."

    The ensemble weight α is the only free parameter of Eq. (1) and is fit to the validation set; DistilBERT is also fine-tuned with performance reported on the same validation set (84.3%). Table I then reports DistilBERT-only accuracy as 0.843, matching that validation number, and the hybrid accuracy as 0.876, with no described held-out test split in Sec. IV-A1. The claimed 'prediction' of 87.6% is therefore an evaluation on the data used to select α and tune the model, not an independent generalization result. The CSI values in Eq. (2) are averages of these same tuned scores, so the corporate findings inherit the fitted evaluation.

full rationale

The central performance claim is not derived from independent evidence: the ensemble weight α is tuned by grid search on the validation set, and the reported Table I numbers appear to be validation-set numbers (DistilBERT-only 0.843 equals the stated validation accuracy). The paper never describes a held-out test split, so the 87.6% accuracy is a fitted in-sample result presented as a measured outcome. This is a clear fitted-input-called-prediction pattern, giving a circularity score of 6 rather than a lower score. The separate internal inconsistency in Eq. (1) — where S_DistilBERT uses the probability of the highest sentiment class without using the class label, so confident negative tweets can produce high scores — is a correctness defect, not a circularity, and is not scored here. The VADER and DistilBERT citations are standard external references, not load-bearing self-citations, so no self-citation circularity is present.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central performance claim rests on one fitted scalar α=0.4 plus hand-chosen thresholds and data-selection cutoffs. The reliability of the manual labels, the VADER/DistilBERT models, and the snscrape sample are treated as external ground truth without independent checks. No new entities are introduced.

free parameters (4)
  • Ensemble weight α = 0.4
    Chosen by grid search on the validation set (Sec III-C3). Since the same validation set appears to be used for the reported accuracy, the performance claim is partly a fit to the evaluation data.
  • Sentiment classification thresholds = Positive S≥0.6, Negative S≤0.4
    Chosen by hand (Sec III-C3) with no sensitivity analysis or justification.
  • Minimum engagement threshold = At least 5 likes or replies
    Applied during scraping (Sec III-A) to filter tweets; this biases the sample toward already-viral posts and is not varied or justified.
  • Tweets per company = 500
    Arbitrary count in Sec III-A used to compute the Corporate Sentiment Index; no power analysis or confidence intervals are provided.
assumptions (4)
  • domain assumption VADER and DistilBERT, with the paper's unspecified modifications, produce valid sentiment scores for X posts.
    The whole pipeline treats both models as reliable, but the exact modifications to VADER (enhanced emoji lexicon, added slang) are not specified or evaluated.
  • domain assumption The 2,000 manually labeled tweets are accurate ground truth for the evaluation.
    No inter-annotator agreement or labeling protocol is reported (Sec IV-A1).
  • domain assumption A held-out test set exists and is distinct from the fine-tuning and validation data.
    The paper never describes a train/validation/test split; sections III-C2 and IV-A1 imply a single manual-label pool, making this independence premise load-bearing.
  • domain assumption The snscrape scrape with the stated filters is representative of public opinion about each company.
    The filters exclude retweets and low-engagement posts (Sec III-A), and the limitations section admits incomplete capture during high-volume events.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Visualizing Public Opinion on X: A Real-Time Sentiment Dashboard Using VADER and DistilBERT." pith.science (2026). https://pith.science/paper/NNOI5AHX

@misc{pith2026250415448,
  author       = {Pith},
  title        = {Pith review of: Visualizing Public Opinion on X: A Real-Time Sentiment Dashboard Using VADER and DistilBERT},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NNOI5AHX}},
  note         = {Machine review of arXiv:2504.15448}
}
read the original abstract

In the age of social media, understanding public sentiment toward major corporations is crucial for investors, policymakers, and researchers. This paper presents a comprehensive sentiment analysis system tailored for corporate reputation monitoring, combining Natural Language Processing (NLP) and machine learning techniques to accurately interpret public opinion in real time. The methodology integrates a hybrid sentiment detection framework leveraging both rule-based models (VADER) and transformer-based deep learning models (DistilBERT), applied to social media data from multiple platforms. The system begins with robust preprocessing involving noise removal and text normalization, followed by sentiment classification using an ensemble approach to ensure both interpretability and contextual accuracy. Results are visualized through sentiment distribution plots, comparative analyses, and temporal sentiment trends for enhanced interpretability. Our analysis reveals significant disparities in public sentiment across major corporations, with companies like Amazon (81.2) and Samsung (45.8) receiving excellent sentiment scores, while Microsoft (21.7) and Walmart (21.9) exhibit poor sentiment profiles. These findings demonstrate the utility of our multi-source sentiment framework in providing actionable insights regarding corporate public perception, enabling stakeholders to make informed strategic decisions based on comprehensive sentiment analysis.

Figures

Figures reproduced from arXiv: 2504.15448 by the authors.

Figure 1
Figure 1. Text preprocessing pipeline for social media content. The process [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Corporate Sentiment Index (CSI) across 20 major companies, showing significant variations in public sentiment from highly positive to predominantly [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 49 canonical work pages

  1. [1]

    Advancements in Sentiment Analysis: A Hybrid Model Approach for Enhanced Accuracy and Interpretability,

    S. Kumar and S. Sharma, “Advancements in Sentiment Analysis: A Hybrid Model Approach for Enhanced Accuracy and Interpretability,” inProc. Data Analytics and Management (ICDAM 2024), pp. 515–528, Springer, 2025

    work page 2024

  2. [2]

    DistilRoBiLSTMFuse: An Efficient Hybrid Deep Learning Approach for Sentiment Analysis,

    S. K. Papia, M. A. Khan, T. Habib, M. Rahman, and M. N. Islam, “DistilRoBiLSTMFuse: An Efficient Hybrid Deep Learning Approach for Sentiment Analysis,”PeerJ Computer Science, vol. 10, e2349, 2024

    work page 2024

  3. [3]

    Exploring Sentiment Analysis on Social Media Texts,

    N. A. Alabdulkarim, M. A. Haq, and J. Gyani, “Exploring Sentiment Analysis on Social Media Texts,”Engineering, Technology & Applied Science Research, vol. 14, no. 3, pp. 14442–14450, 2024

    work page 2024

  4. [4]

    RoBERTa-BiLSTM: A Context-Aware Hybrid Model for Sentiment Analysis

    M. M. Rahman, A. I. Shiplu, Y . Watanobe, and M. A. Alam, “RoBERTa- BiLSTM: A Context-Aware Hybrid Model for Sentiment Analysis,” arXiv preprint arXiv:2406.00367, 2024

    work page Pith review arXiv 2024

  5. [5]

    Faces Speak Louder Than Words: Emotions Versus Textual Sentiment in the 2024 USA Presidential Election

    C. Wei, S. Noh, and H. H. Chang, “Faces Speak Louder Than Words: Emotions Versus Textual Sentiment in the 2024 USA Presidential Election,” arXiv preprint arXiv:2412.18031, 2024

    work page Pith review arXiv 2024

  6. [6]

    A Hybrid Approach to Aspect-Based Sentiment Analysis Using Transfer Learning,

    G. Negi, R. Sarkar, O. Zayed, and P. Buitelaar, “A Hybrid Approach to Aspect-Based Sentiment Analysis Using Transfer Learning,” arXiv preprint arXiv:2403.17254, 2024

    arXiv 2024

  7. [7]

    SentiQNF: A Novel Approach to Sentiment Analysis Using Quantum Algorithms and Neuro-Fuzzy Systems

    K. Dave, N. Innan, B. K. Behera, Z. Mumtaz, S. Al-Kuwari, and A. Farouk, “SentiQNF: A Novel Approach to Sentiment Analysis Using Quantum Algorithms and Neuro-Fuzzy Systems,” arXiv preprint arXiv:2412.12731, 2024

    work page Pith review arXiv 2024

  8. [8]

    Social Media Sentiment Analysis for Brand Reputation Management,

    M. Hammad, N. A. Semary, K. M. Amin, and W. Ahmed, “Social Media Sentiment Analysis for Brand Reputation Management,”SSRN Electronic Journal, 2024

    work page 2024

Show all 52 references
  1. [9]

    An Efficient Hybrid Deep Learning Approach for Sentiment Analysis,

    S. Khan, M. A. Khan, and T. Habib, “An Efficient Hybrid Deep Learning Approach for Sentiment Analysis,”PeerJ Computer Science, vol. 10, e2349, 2024

    2024

  2. [10]

    Aspect-Level Sentiment Analysis of Customer Reviews Based on Neural Multi-task Learning,

    Y . Shi, L. Li, H. Li, A. Li, and Y . Lin, “Aspect-Level Sentiment Analysis of Customer Reviews Based on Neural Multi-task Learning,”J. Theory Pract. Eng. Sci., vol. 4, no. 4, pp. 1–8, 2024

    2024

  3. [11]

    Dynamic Sentiment Adaptation in Streaming Social Data,

    H. Chen and Q. Zhang, “Dynamic Sentiment Adaptation in Streaming Social Data,”Data Mining and Knowledge Discovery, vol. 38, no. 1, pp. 45–67, 2024

    2024

  4. [12]

    AI-Driven Reputation Management in Global Corporations,

    McKinsey & Company, “AI-Driven Reputation Management in Global Corporations,” Tech. Rep., McKinsey Digital, 2024

    2024

  5. [13]

    Transformers Library for Real-Time NLP,

    Hugging Face Team, “Transformers Library for Real-Time NLP,” Ver. 4.25, 2024

    2024

  6. [14]

    Real-Time Twitter Sentiment Corpus,

    Twitter Developer Platform, “Real-Time Twitter Sentiment Corpus,” Dataset, 2024. Available: https://developer.twitter.com/en/docs

    2024

  7. [15]

    Comparative Analysis of Corporate Reputa- tion Management Tools,

    V orecol Analytics Team, “Comparative Analysis of Corporate Reputa- tion Management Tools,” Tech. Rep., V orecol HRMS, 2024. Available: https://vorecol.com/blogs/corporate-reputation-2023

    2024

  8. [16]

    Sentimental Analysis and Social Media,

    TechRxiv, “Sentimental Analysis and Social Media,” 2024

    2024

  9. [17]

    Feasible Sentiment Analysis of Real-Time Twitter Data,

    A. Sharma and B. Gupta, “Feasible Sentiment Analysis of Real-Time Twitter Data,”E3S Web of Conferences, vol. 430, p. 01045, 2023

    2023

  10. [18]

    Multimodal Sentiment Analysis: A Comprehensive Survey,

    Y . Liu, C. Huang, and L. Wang, “Multimodal Sentiment Analysis: A Comprehensive Survey,” arXiv preprint arXiv:2305.07611, 2023

    2023 arXiv

  11. [19]

    Deep Neural Architectures for Social Media Sentiment Mining,

    K. Garg and A. Kumar, “Deep Neural Architectures for Social Media Sentiment Mining,”IEEE Transactions on Computational Social Sys- tems, vol. 10, no. 2, pp. 112–125, 2023

    2023

  12. [20]

    2023 Trust Barometer: Social Media Sentiment Trends,

    Edelman Intelligence, “2023 Trust Barometer: Social Media Sentiment Trends,”Edelman Trust Report, pp. 22–45, 2023

    2023

  13. [21]

    A Hybrid Model for Sentiment Analysis Based on Movie Review Classification,

    R. Kumar and P. Singh, “A Hybrid Model for Sentiment Analysis Based on Movie Review Classification,” inProc. Int. Conf. on Recent Trends in Computing, pp. 424–431, 2023

    2023

  14. [22]

    Attention-Based Hybrid Models for Real-Time Sentiment Tracking,

    Y . Wang and L. Zhao, “Attention-Based Hybrid Models for Real-Time Sentiment Tracking,” inProc. ACM Web Conf., pp. 234–245, 2023

    2023

  15. [23]

    Sentiment Analysis of Russia-Ukraine Con- flict Tweets Using RoBERTa,

    L. Ramos and O. Chang, “Sentiment Analysis of Russia-Ukraine Con- flict Tweets Using RoBERTa,”Uniciencia, vol. 37, no. 1, pp. 421–431, 2023

    2023

  16. [24]

    An Ensemble Multi-Layered Sentiment Anal- ysis Model (EMLSA),

    R.K. Singh and A. Patel, “An Ensemble Multi-Layered Sentiment Anal- ysis Model (EMLSA),” inFuture of Information and Communication Conference, pp. 1–15, 2023

    2023

  17. [25]

    A Review on Sentiment Analysis from Social Media Platforms,

    M. Rodr ´ıguez-Ib´a˜nez, A. Cas ´anez-Ventura, F. Castej´on-Mateos, and P.- M. Cuenca-Jim ´enez, “A Review on Sentiment Analysis from Social Media Platforms,”J. Artif. Intell. Res., vol. 45, pp. 112–145, 2023

    2023

  18. [26]

    Evaluating SA on Twitter to inform investment decisions,

    Hasselgren et al., “Evaluating SA on Twitter to inform investment decisions,” 2023

    2023

  19. [27]

    Sentiment Analysis: Trends and Applications,

    Zhang et al., “Sentiment Analysis: Trends and Applications,” 2023

    2023

  20. [28]

    BERTweet: A pre-trained language model for English Tweets,

    D.Q. Nguyen, “BERTweet: A pre-trained language model for English Tweets,” 2023. Available: https://github.com/VinAIResearch/BERTweet

    2023

  21. [29]

    Sentiment Analysis in Various Domains,

    Rodr ´ıguez-Ib´a˜nez et al., “Sentiment Analysis in Various Domains,” 2023

    2023

  22. [30]

    The web as a significant and universal information source,

    Birjali et al., “The web as a significant and universal information source,” 2021

    2021

  23. [31]

    A Text Preprocessing Toolkit for Topic Mod- eling on Social Media Data,

    R. Churchill and L. Singh, “A Text Preprocessing Toolkit for Topic Mod- eling on Social Media Data,” inProc. 13th Int. Joint Conf. on Knowledge Discovery, Knowledge Engineering and Knowledge Management, 2021

    2021

  24. [32]

    Decision support from financial disclo- sures with deep neural networks and transfer learning,

    M. Kraus and S. Feuerriegel, “Decision support from financial disclo- sures with deep neural networks and transfer learning,”Decision Support Systems, vol. 135, p. 113332, 2020

    2020

  25. [33]

    Attention-based LSTM for aspect-level sentiment classification,

    Y . Wang, M. Huang, X. Zhu, and L. Zhao, “Attention-based LSTM for aspect-level sentiment classification,” inProc. EMNLP, pp. 606–615, 2020

    2020

  26. [34]

    How to fine-tune BERT for text classification?,

    C. Sun, X. Qiu, Y . Xu, and X. Huang, “How to fine-tune BERT for text classification?,” inChinese Computational Linguistics, 2019, pp. 194–206

    2019

  27. [35]

    Assessing state-of-the- art sentiment models on state-of-the-art sentiment datasets,

    J. Barnes, R. Klinger, and S. Schulte im Walde, “Assessing state-of-the- art sentiment models on state-of-the-art sentiment datasets,” inProc. 8th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis, 2019, pp. 2–12

    2019

  28. [36]

    BERT: Pre-training of deep bidirectional transformers for language understanding,

    J. Devlin, M.W. Chang, K. Lee, and K. Toutanova, “BERT: Pre-training of deep bidirectional transformers for language understanding,” inProc. 2019 Conf. of the NAACL: HLT, pp. 4171–4186, 2019

    2019

  29. [37]

    DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter,

    V . Sanh, L. Debut, J. Chaumond, and T. Wolf, “DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter,” arXiv preprint arXiv:1910.01108, 2019

    1910 arXiv

  30. [38]

    Social sentiment sensor in Twitter for predicting cyber-attacks using ℓ1 regularization,

    A. Hernandez-Suarez, G. Sanchez-Perez, K. Toscano-Medina, V . Martinez-Hernandez, H. Perez-Meana, and J. Olivares-Mercado, “Social sentiment sensor in Twitter for predicting cyber-attacks using ℓ1 regularization,”Sensors, vol. 18, no. 5, p. 1380, 2018

    2018

  31. [39]

    A three-stage methodology for SA on big social data,

    El Alaoui et al., “A three-stage methodology for SA on big social data,” 2018

    2018

  32. [40]

    Brand-related Twitter sentiment analysis using feature engineering and the dynamic architecture for artificial neural networks,

    D. Zimbra, M. Ghiassi, and S. Lee, “Brand-related Twitter sentiment analysis using feature engineering and the dynamic architecture for artificial neural networks,” inProc. 51st Hawaii Int. Conf. on System Sciences, pp. 1909–1918, 2018

    1909

  33. [41]

    Enhancing deep learning sentiment analysis with ensemble techniques in social applications,

    O. Araque, I. Corcuera-Platas, J.F. S ´anchez-Rada, and C.A. Iglesias, “Enhancing deep learning sentiment analysis with ensemble techniques in social applications,”Expert Systems with Applications, vol. 77, pp. 236–246, 2017

    2017

  34. [42]

    Twitter sentiment classification using distant supervision,

    A. Go, R. Bhayani, and L. Huang, “Twitter sentiment classification using distant supervision,” Stanford University, Tech. Rep., 2009

    2009

  35. [43]

    Global Vectors for Word Representation,

    J. Pennington and R. Socher, “Global Vectors for Word Representation,”

  36. [44]

    V ADER Sentiment Analysis,

    C.J. Hutto, “V ADER Sentiment Analysis,” GitHub repository, https:// github.com/cjhutto/vaderSentiment

  37. [45]

    V ADER sentiment analysis (with examples),

    Hex, “V ADER sentiment analysis (with examples),” 2023. [Online]. Available: https://hex.tech/templates/sentiment-analysis/ vader-sentiment-analysis/

    2023

  38. [46]

    V ADER: A parsimonious rule-based model for sentiment analysis of social media text,

    C.J. Hutto and E.E. Gilbert, “V ADER: A parsimonious rule-based model for sentiment analysis of social media text,” inProc. 8th Int. AAAI Conf. on Weblogs and Social Media, 2014, pp. 216–225

    2014

  39. [47]

    What is Data Preprocessing? Key Steps and Techniques,

    TechTarget, “What is Data Preprocessing? Key Steps and Techniques,” 2022

    2022

  40. [48]

    Text Mining and Pre-Processing Methods for Social Media Data Extraction and Processing,

    IGI Global, “Text Mining and Pre-Processing Methods for Social Media Data Extraction and Processing,” 2022

    2022

  41. [49]

    Improving SM SA by merging social indicators with textual content,

    Ilk & Fan, “Improving SM SA by merging social indicators with textual content,” 2022

    2022

  42. [50]

    A review and theoretical analysis of data mining techniques for social media sentiment analysis,

    A. Kumar, K. Garg, G. Kaur, and G.P. Garg, “A review and theoretical analysis of data mining techniques for social media sentiment analysis,” Int. J. Inf. Manage. Data Insights, vol. 2, no. 1, p. 100059, 2022

    2022

  43. [51]

    Attention-based LSTM network for cross-lingual sentiment classification,

    X. Zhou, X. Wan, and J. Xiao, “Attention-based LSTM network for cross-lingual sentiment classification,” inProc. EMNLP, pp. 247–256, 2020

    2020

  44. [2024]

    Available: https://nlp.stanford.edu/projects/glove/

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.