pith. sign in

arxiv: 2606.20751 · v1 · pith:V2JX2I3Enew · submitted 2026-06-18 · 💻 cs.CL

From Sentiment to Actionable Insights: A Data-Driven Public Sentiment Analysis of Advanced Air Mobility

Esrat Farhana Dulia , Amina Dhaher , Raiful Hasan , Syed Arbab Mohd Shihab This is my paper

Pith reviewed 2026-06-26 17:50 UTC · model grok-4.3

classification 💻 cs.CL
keywords Advanced Air Mobilitypublic sentiment analysissentiment classificationtopic modelingLDAModernBERTpublic acceptancedrone technology
0
0 comments X

The pith

Analysis of over 300,000 online texts identifies six clusters of public concern that shape acceptance of advanced air mobility.

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

The paper evaluates seven sentiment models on AAM-related posts from Reddit and Quora, selects ModernBERT as the strongest performer, and applies it to label the full collection of 306,009 texts. It then runs LDA within each sentiment class to surface 20 topics whose distribution across positive and negative classes reveals six dominant clusters of public worry. A sympathetic reader would care because these clusters directly affect the regulatory, commercial, and operational path for AAM systems whose success depends on public support.

Core claim

ModernBERT outperforms the other six models on AAM-specific classification; when its labels are used to drive per-sentiment LDA, the resulting topics collapse into six clusters whose shares are workforce and skill development (25.29 percent), regulation and compliance (24.64 percent), technical performance of drones (20.99 percent), military/geopolitics/defense (14.58 percent), safety and operational risks (8.51 percent), and noise and disturbance (5.98 percent).

What carries the argument

ModernBERT sentiment classifier feeding per-class Latent Dirichlet Allocation to extract and cross-tabulate 20 latent topics from 306,009 texts.

If this is right

  • Targeted interventions on workforce training and regulatory clarity could address the two largest clusters and raise overall acceptance.
  • Tracking topic evolution from 2008 to 2025 supplies a baseline for monitoring how public discourse shifts with real-world AAM deployments.
  • Cross-sentiment topic breakdowns allow operators to distinguish concerns that appear mainly in negative posts from those that also appear in positive ones.
  • The six-cluster map supplies a concrete checklist for government and industry communication strategies.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same pipeline could be rerun on future data to test whether cluster shares change after regulatory milestones or first commercial flights.
  • Comparing these online-derived clusters against structured survey instruments would quantify any selection bias in Reddit and Quora sources.
  • Extending the method to other emerging transport technologies would show whether the same six concern types recur across domains.

Load-bearing premise

The collected Reddit and Quora texts form a representative sample of wider public opinion on AAM and that off-the-shelf models trained on general text transfer accurately to AAM language without domain-specific retraining or ground-truth validation.

What would settle it

A probability-sampled national survey that reports materially different shares for the six clusters or surfaces additional clusters not captured in the online corpus.

Figures

Figures reproduced from arXiv: 2606.20751 by Amina Dhaher, Esrat Farhana Dulia, Raiful Hasan, Syed Arbab Mohd Shihab.

Figure 1
Figure 1. Figure 1: Overview of the methodology for AAM sentiment analysis. First Author et al.: Preprint submitted to Elsevier Page 39 of 38 [PITH_FULL_IMAGE:figures/full_fig_p039_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Number of texts in each sentiment class (Neutral, Positive, Negative, and Others) in the sample dataset before back translation and after back translation (final annotated dataset used for model fine-tuning). First Author et al.: Preprint submitted to Elsevier Page 40 of 38 [PITH_FULL_IMAGE:figures/full_fig_p040_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Confusion matrix of ModernBERT predictions on the test dataset [PITH_FULL_IMAGE:figures/full_fig_p041_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Sentiment distribution of the full AAM dataset (301,459 unannotated texts) predicted by ModernBERT. First Author et al.: Preprint submitted to Elsevier Page 41 of 38 [PITH_FULL_IMAGE:figures/full_fig_p041_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Distribution of net vote scores (upvotes minus downvotes) across predicted sentiment classes [PITH_FULL_IMAGE:figures/full_fig_p042_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Temporal distribution of the six topics within the Positive sentiment class from 2008 to mid-2025. Topic 4 (Drone Hardware and Flight Control) and Topic 3 (Personal Experiences of Using Drones) dominate throughout most of the study period. Topic 1 (Drone Operations and Applications) shows notable growth after 2020, while Topic 6 (Drone Delivery and Future Urban Mobility) gradually increases over time. Topi… view at source ↗
Figure 7
Figure 7. Figure 7: Temporal trends of negative drone-related discussions (2008–mid-2025). Topic 3 (Regulation and Compliance) and Topic 1 (Safety Concerns) are the most discussed overall. Topic 6 (Unethical Use od Drones) shows a sharp spike in 2024 driven by geopolitical conflicts. Topic 2 (Noise Concerns and Perceived Disturbance) remains stable over time, Topic 5 (Technical Constraints) shows a steady mid-level presence, … view at source ↗
Figure 8
Figure 8. Figure 8: Temporal distribution of the four topics within the Neutral sentiment class from 2008 to mid-2025. All topics exhibit noticeable increases between 2012 and 2016, followed by another rise after 2020. Topic 1 (Drone Industry, Workforce, and Production) remains the most consistently dominant topic throughout the study period. Topic 2 (Drone Regulations and Airspace Management) increases sharply in recent year… view at source ↗
Figure 9
Figure 9. Figure 9: Temporal distribution of the four topics within the Others sentiment class from 2008 to mid-2025. All topics show a noticeable increase around 2015, followed by another rise after 2020. Topic 3 (Energy Systems of Drone) is the dominant topic for most of the study period, particularly from 2010 to 2022, consistently recording higher counts than the other topics. Topic 2 (Drone Rules, Pilot Guidance, and Reg… view at source ↗
Figure 10
Figure 10. Figure 10: Topic distributions across the four sentiment classes, showing the percentage of texts from the 301,459 texts associated with each topic within each class. The Positive and Negative sentiment classes each contain six topics, while the Neutral and Others classes each contain four topics. In the Positive class, Topics 3 (Personal Experiences of Using Drones) and 4 (Drone Hardware and Flight Control) account… view at source ↗
Figure 11
Figure 11. Figure 11: Yearly trends of sentiment-labeled texts across six clusters of concern derived from topic aggregation. Each subplot shows the temporal distribution of Negative, Neutral, and Others sentiment classes for a specific cluster: (a) Regulation and Compliance, (b) Safety and Operational Risks, (c) Technical Performance of Drones, (d) Military, Geopolitics, and Defense, (e) Workforce, Skill Development, and Care… view at source ↗
Figure 12
Figure 12. Figure 12: Distribution of AAM discussions across identified clusters of concern. Each cluster represents a proportion of the entire dataset (all texts across all clusters and sentiment classes), while the stacked bars show the internal sentiment composition (Negative, Neutral, and Others) within each cluster. The total height of each bar indicates the number of texts in that cluster, and the internal segments repre… view at source ↗
read the original abstract

Advanced Air Mobility (AAM) is an emerging low-altitude air transportation system whose successful deployment depends not only on technological advancement but also on public acceptance. This acceptance will drive government support, regulations, noise standards, and willingness to fly, and in turn the overall commercial viability of AAM. Understanding public sentiment toward AAM is therefore essential for identifying its societal barriers and informing its adoption strategies. This study analyzes 306,009 human-generated texts collected from Reddit and Quora to examine public discourse on AAM using AI-based models. Because multiple sentiment analysis models exist, identifying the most accurate model is critical for reliable AAM sentiment prediction and trustworthy public opinion analysis. Accordingly, seven models spanning lexicon-based, machine learning, deep learning, and transformer-based approaches are evaluated for AAM-specific sentiment classification. ModernBERT achieves the best classification performance and is used to label the full dataset. Using the resulting sentiment labels, Latent Dirichlet Allocation (LDA) is applied within each sentiment class to uncover latent topics in public opinion. The analysis identifies 20 distinct topics and traces their temporal evolution from 2008 to 2025. A cross-sentiment topic analysis further reveals six major clusters of public concern: workforce and skill development (25.29% of the dataset), regulation and compliance (24.64%), technical performance of drones (20.99%), military, geopolitics, and defense (14.58%), safety and operational risks (8.51%), and noise and disturbance (5.98%). Based on these findings, this study provides actionable strategies to address these concerns, thereby, improving public acceptance and support AAM deployment.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 1 minor

Summary. The manuscript analyzes 306,009 Reddit and Quora texts on Advanced Air Mobility (AAM). It evaluates seven sentiment models (lexicon-based through transformer-based) for AAM-specific classification, selects ModernBERT as best, labels the full corpus, applies LDA within each sentiment class to extract 20 topics, traces their evolution from 2008–2025, and reports six cross-sentiment clusters with exact percentages (workforce/skill development 25.29%, regulation/compliance 24.64%, technical performance 20.99%, military/geopolitics/defense 14.58%, safety/operational risks 8.51%, noise/disturbance 5.98%). Actionable strategies for improving public acceptance are derived from the clusters.

Significance. If the sentiment labels prove reliable on AAM text, the scale of the corpus and the temporal plus cross-sentiment topic analysis would supply concrete, policy-relevant evidence on societal barriers to AAM deployment. The explicit cluster percentages and the 2008–2025 trend analysis are potentially useful for regulators and industry.

major comments (3)
  1. [Abstract] Abstract and model-evaluation section: the claim that seven models were evaluated for AAM-specific sentiment classification and that ModernBERT was selected is unsupported by any performance numbers, train/test splits, error bars, or comparison against an AAM-held-out set. General-corpus accuracy does not establish transfer to AAM terminology (regulation, noise, workforce, geopolitics), rendering the downstream LDA topics and six-cluster percentages unreliable.
  2. [Results (topic clusters)] Topic-modeling and results sections: the reported cluster shares (e.g., workforce 25.29%, regulation 24.64%) and the 20-topic solution rest entirely on the unvalidated ModernBERT labels. Without a human-annotated AAM validation set, any reported accuracy on non-AAM data cannot confirm that the clusters capture genuine public concerns rather than label noise.
  3. [Methods] Methods (data and labeling pipeline): the 306,009 Reddit/Quora posts are treated as a representative sample of broader public sentiment, yet no domain-specific fine-tuning or ground-truth AAM labels are described. This assumption is load-bearing for all temporal-evolution and cross-sentiment claims.
minor comments (1)
  1. [Abstract] The abstract states the temporal range extends to 2025; the manuscript should explicitly state how posts dated after the current date were obtained or whether the range is a projection.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We agree that the model evaluation and validation steps require substantially more documentation and empirical support to ensure the reliability of the sentiment labels and downstream topic clusters. We will revise the manuscript to address each of these points.

read point-by-point responses

  1. Referee: [Abstract] Abstract and model-evaluation section: the claim that seven models were evaluated for AAM-specific sentiment classification and that ModernBERT was selected is unsupported by any performance numbers, train/test splits, error bars, or comparison against an AAM-held-out set. General-corpus accuracy does not establish transfer to AAM terminology (regulation, noise, workforce, geopolitics), rendering the downstream LDA topics and six-cluster percentages unreliable.

    Authors: We agree that the current manuscript does not provide sufficient detail on the model evaluation. In the revised version we will add a dedicated model-evaluation subsection containing a comparison table for all seven models. The table will report accuracy, macro-F1, and other metrics on both general corpora and an AAM-specific held-out set (with train/test splits and confidence intervals). This will explicitly justify the selection of ModernBERT for the AAM domain and support the subsequent LDA and clustering results. revision: yes

  2. Referee: [Results (topic clusters)] Topic-modeling and results sections: the reported cluster shares (e.g., workforce 25.29%, regulation 24.64%) and the 20-topic solution rest entirely on the unvalidated ModernBERT labels. Without a human-annotated AAM validation set, any reported accuracy on non-AAM data cannot confirm that the clusters capture genuine public concerns rather than label noise.

    Authors: The referee is correct that the cluster percentages depend on the quality of the ModernBERT labels. We will add a human-validation experiment: a random sample of AAM posts will be independently annotated by multiple annotators for sentiment; we will report inter-annotator agreement and the agreement between ModernBERT predictions and human labels. These results will be presented before the LDA and cross-sentiment clustering sections so that readers can assess the reliability of the reported topic shares. revision: yes

  3. Referee: [Methods] Methods (data and labeling pipeline): the 306,009 Reddit/Quora posts are treated as a representative sample of broader public sentiment, yet no domain-specific fine-tuning or ground-truth AAM labels are described. This assumption is load-bearing for all temporal-evolution and cross-sentiment claims.

    Authors: We will revise the methods section to state explicitly that no domain-specific fine-tuning was performed and that the seven models were applied using their publicly released weights. A new limitations paragraph will discuss the representativeness of Reddit and Quora data, the lack of ground-truth AAM sentiment labels, and the implications for temporal and cross-sentiment analyses. We will also note that the scale of the corpus still provides useful evidence of online discourse even if it is not a perfect proxy for the general population. revision: yes

Circularity Check

0 steps flagged

Standard empirical pipeline with no circular derivation

full rationale

The paper describes a standard empirical workflow: evaluate seven off-the-shelf sentiment models on AAM-related text, select the best performer (ModernBERT), apply it to label 306k posts, then run LDA topic modeling and cluster analysis. No equations, fitted parameters, self-citations, or uniqueness theorems appear in the provided text that would reduce any claimed result to an input by construction. The central outputs (topic percentages, temporal trends) are downstream applications of the labeling step rather than self-referential derivations, satisfying the criteria for a non-circular empirical study.

Axiom & Free-Parameter Ledger

1 free parameters · 0 axioms · 0 invented entities

The analysis rests on standard NLP assumptions and the representativeness of social-media text; no new free parameters beyond the conventional choice of 20 topics are introduced in the abstract, and no invented entities are postulated.

free parameters (1)
  • number of topics
    LDA run within each sentiment class; abstract states 20 topics were identified but does not report how this number was chosen or validated.

pith-pipeline@v0.9.1-grok · 5844 in / 1323 out tokens · 18570 ms · 2026-06-26T17:50:17.943163+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

109 extracted references · 87 canonical work pages · 3 internal anchors

  1. [1]

    Building a Performance Comparison Framework for Urban Air Mobility Airspace Management Concepts, in: 2023 IEEE/AIAA 42nd Digital Avionics Systems Conference (DASC), pp. 1–10. doi:10.1109/DASC58513.2023.10311113. Aeronautics, N., Administration, S.,

    work page doi:10.1109/dasc58513.2023.10311113 2023

  2. [2]

    Available online: https://www.nasa.gov/aam

    Advanced Air Mobility (AAM) Mission Overview. Available online: https://www.nasa.gov/aam. Ahmed,H.U.,Huang,Y.,Lu,P.,Bridgelall,R.,2022. Technologydevelopmentsandimpactsofconnectedandautonomousvehicles:Anoverview. Smart Cities 5, 382–404. doi:10.3390/smartcities5010022. Al Haddad, C., Chaniotakis, E., Straubinger, A., Plötner, K.O., Antoniou, C.,

    work page doi:10.3390/smartcities5010022 2022

  3. [3]

    Transportation Research Part A: Policy and Practice 132, 696–712

    Factors affecting the adoption and use of urban air mobility. Transportation Research Part A: Policy and Practice 132, 696–712. doi:10.1016/j.tra.2019.12.020. Albalate, D., Bel, G.,

    work page doi:10.1016/j.tra.2019.12.020 2019

  4. [4]

    Public administration review 72, 336–349

    High-speed rail: Lessons for policy makers from experiences abroad. Public administration review 72, 336–349. doi:10.1111/j.1540-6210.2011.02492.x. Altun, A.T., Hasanzade, M., Saldiran, E., Guner, G., Uzun, M., Fremond, R., Tang, Y., Bhundoo, P., Su, Y., Xu, Y., et al.,

    work page doi:10.1111/j.1540-6210.2011.02492.x 2011

  5. [5]

    Amelia,P.N.,Atiqah,N.B.,Hasibuan,A.,2026.PublicsentimentanalysisofgovernmentsubsidypoliciesonTwitterusingtheNaïveBayesclassifier

    doi:10.3390/aerospace10080712. Amelia,P.N.,Atiqah,N.B.,Hasibuan,A.,2026.PublicsentimentanalysisofgovernmentsubsidypoliciesonTwitterusingtheNaïveBayesclassifier. Journal of Intelligent Computing and Advanced Data Science 1, 1–15. Aposporis, P.,

    work page doi:10.3390/aerospace10080712 2026

  6. [6]

    Transportation Research Interdisciplinary Perspectives 24, 101064

    A review of global and regional frameworks for the integration of an unmanned aircraft system in air traffic management. Transportation Research Interdisciplinary Perspectives 24, 101064. doi:https://doi.org/10.1016/j.trip.2024.101064. Artstein, R.,

    work page doi:10.1016/j.trip.2024.101064 2024

  7. [7]

    URL:https://doi.org/10.1007/978-94-024-0881-2_11, doi:10.1007/ 978-94-024-0881-2_11

    Inter-annotator agreement. URL:https://doi.org/10.1007/978-94-024-0881-2_11, doi:10.1007/ 978-94-024-0881-2_11. Artstein, R., Poesio, M.,

    work page doi:10.1007/978-94-024-0881-2_11

  8. [8]

    Technology in society 59, 101180

    Public acceptance of drones: Knowledge, attitudes, and practice. Technology in society 59, 101180. doi:10.1016/j.techsoc. 2019.101180. Bahdanau, D., Cho, K., Bengio, Y.,

    work page doi:10.1016/j.techsoc 2019

  9. [9]

    Bansal,P.,Kockelman,K.,Singh,A.,2016

    doi:doi.org/10.3390/logistics4020012. Bansal,P.,Kockelman,K.,Singh,A.,2016. AssessingPublicOpinionsofandInterestinNewVehicleTechnologies. TransportationResearchPart C doi:10.1016/j.trc.2016.01.019. Barari, G., Janke, C., Lin, Y.,

    work page doi:10.3390/logistics4020012 2016

  10. [10]

    Batool,A.,Byun,Y.C.,2024.EnhancedSentimentAnalysisandTopicModelingDuringthePandemicUsingAutomatedLatentDirichletAllocation

    doi:10.3390/admsci16010053. Batool,A.,Byun,Y.C.,2024.EnhancedSentimentAnalysisandTopicModelingDuringthePandemicUsingAutomatedLatentDirichletAllocation. IEEE Access 12, 81206–81220. doi:10.1109/ACCESS.2024.3411717. Bauranov, A., Rakas, J.,

    work page doi:10.3390/admsci16010053 2024

  11. [11]

    Progress in Aerospace Sciences 125, 100726

    Designing airspace for urban air mobility: A review of concepts and approaches. Progress in Aerospace Sciences 125, 100726. doi:10.1016/j.paerosci.2021.100726. Bengio, Y.,

    work page doi:10.1016/j.paerosci.2021.100726 2021

  12. [12]

    Foundations and Trends in Machine Learning 2, 1–127

    Learning deep architectures for ai. Foundations and Trends in Machine Learning 2, 1–127. doi:10.1561/2200000006. Bengio, Y., Simard, P., Frasconi, P.,

    work page doi:10.1561/2200000006

  13. [13]

    IEEE Transactions on Neural Networks 5, 157–166

    Learning long-term dependencies with gradient descent is difficult. IEEE Transactions on Neural Networks 5, 157–166. doi:10.1109/72.279181. Bhaduri, E., Choudhury, C.F., 2026a. Shifting skies: A cross-country investigation of evolution of public perception toward urban air mobility through twitter (x) discourse. Journal of Air Transport Management 132, 10...

    work page doi:10.1109/72.279181 2025

  14. [14]

    Journal of Machine Learning Research 3, 993–1022

    Latent Dirichlet Allocation. Journal of Machine Learning Research 3, 993–1022. Butterworth-Hayes,P.,2024. Easapublishescompletedpackageofadvancedairmobilityregulations. URL:https://www.unmannedairspace. info/uncategorized/easa-publishes-completed-package-of-advanced-air-mobility-regulations/. published: May 24, 2024; Accessed: 2026-05-14. Cartile, A., Mar...

    2024

  15. [15]

    Cero, I., Luo, J., Falligant, J.M.,

    doi:10.3390/aerospace12080724. Cero, I., Luo, J., Falligant, J.M.,

    work page doi:10.3390/aerospace12080724

  16. [16]

    Perspectives on behavior science 47, 283–310

    Lexicon-based sentiment analysis in behavioral research. Perspectives on behavior science 47, 283–310. doi:10.1007/s40614-023-00394-x. Chancey,E.T.,Politowicz,M.,2020a.DesigningandTrainingforAppropriateTrustinIncreasinglyAutonomousAdvancedAirMobilityOperations: A Mental Model Approach: Version

    work page doi:10.1007/s40614-023-00394-x

  17. [17]

    National Aeronautics and Space Administration (NASA)

    Technical Memorandum. National Aeronautics and Space Administration (NASA). NASA Document ID: 20205003378. Chancey, E.T., Politowicz, M.S., 2020b. Public trust and acceptance for concepts of remotely operated urban air mobility transportation, in: Proceedings of the Human Factors and Ergonomics Society Annual Meeting, pp. 1516–1520. doi:10.1177/1071181320...

    work page doi:10.1177/1071181320641251 2021

  18. [18]

    Risk analysis 35, 1167–1183

    Risk Perception and the Public Acceptance of Drones. Risk analysis 35, 1167–1183. doi:10.1111/risa.12330. Cohen, A.P., Shaheen, S.A., Farrar, E.M.,

    work page doi:10.1111/risa.12330

  19. [19]

    IEEE Transactions on Intelligent Transportation Systems 22, 6074–6087

    Urban air mobility: History, ecosystem, market potential, and challenges. IEEE Transactions on Intelligent Transportation Systems 22, 6074–6087. doi:10.1109/TITS.2021.3082767. Collobert, R., Weston, J., Bottou, L., Karlen, M., Kavukcuoglu, K., Kuksa, P.,

    work page doi:10.1109/tits.2021.3082767 2021

  20. [20]

    URL:https://adamconnell.me/reddit-statistics/

    Reddit user age, gender, and demographics statistics (2026). URL:https://adamconnell.me/reddit-statistics/. accessed: 2026-05-14. Current Trends,

    2026

  21. [21]

    URL:https://www.currenttrends

    After twitter: How fragmented social media is rebuilding the online public square. URL:https://www.currenttrends. news/2026/01/after-twitter-how-fragmented-social.html. accessed: 2026-05-14. Dällenbach,N.,Wüstenhagen,R.,2022. Howfardonoiseconcernstravel?exploringhowfamiliarityandjusticeshapenoiseexpectationsandsocial acceptanceofplannedwindenergyprojects....

    work page doi:10.1016/j.erss.2021 2026

  22. [22]

    Journal of the American Society for Information Science 41, 391–407

    Indexing by latent semantic analysis. Journal of the American Society for Information Science 41, 391–407. doi:10.1002/(SICI)1097-4571(199009)41:6<391::AID-ASI1>3.0.CO;2-9. Demszky, D., Movshovitz-Attias, D., Ko, J., Cowen, A., Nemade, G., Ravi, S.,

    work page doi:10.1002/(sici)1097-4571(199009)41:6

  23. [23]

    GoEmotions: A Dataset of Fine-Grained Emotions, in: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, Association for Computational Linguistics, Online. pp. 4040–4054. URL:https://www.aclweb.org/anthology/2020.acl-main.372. Deng, S., Sinha, A.P., Zhao, H.,

    2020

  24. [24]

    Decision Support Systems 94, 65–76

    Adapting sentiment lexicons to domain-specific social media texts. Decision Support Systems 94, 65–76. doi:10.1016/j.dss.2016.11.001. Devlin, J., Chang, M.W., Lee, K., Toutanova, K.,

    work page doi:10.1016/j.dss.2016.11.001 2016

  25. [25]

    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding, in: NAACL-HLT, Minneapolis, Minnesota, USA. pp. 4171–4186. DiBattista,A.,Grayling,S.,Hasselaar,E.,Leopold,T.,Li,R.,Rayner,M.,Zahidi,S.,2023. Futureofjobsreport2023,in:WorldEconomicForum, World Economic Forum Geneva, Switzerland, Geneva, Switzerland. pp. 978–2. Dolata, M., Sc...

    2023

  26. [26]

    Government Information Quarterly 40, 101874

    Moving beyond privacy and airspace safety: Guidelines for just drones in policing. Government Information Quarterly 40, 101874. doi:10.1016/j.giq.2023.101874. Doo,J.T.,Pavel,M.D.,Didey,A.,Hange,C.,Diller,N.P.,Tsairides,M.A.,Smith,M.,Bennet,E.,Bromfield,M.,Mooberry,J.,2021. NASAElectric Vertical Takeoff and Landing (eVTOL) Aircraft Technology for Public Se...

    work page doi:10.1016/j.giq.2023.101874 2023

  27. [27]

    Transportation research part A: policy and practice 46, 1241–1251

    The influence of individual’s risk perception and attitudes on travel behavior. Transportation research part A: policy and practice 46, 1241–1251. doi:https://doi.org/10.1016/j.tra.2012.05.013. Elman, J.L.,

    work page doi:10.1016/j.tra.2012.05.013 2012

  28. [28]

    Cognitive Science 14, 179–211

    Finding Structure in Time. Cognitive Science 14, 179–211. doi:10.1207/s15516709cog1402\_1. European Union Aviation Safety Agency (EASA),

    work page doi:10.1207/s15516709cog1402

  29. [29]

    URL:https://www.easa.europa.eu/en/domains/ drones-air-mobility/drones-air-mobility-landscape/urban-air-mobility-uam

    Urban air mobility (uam). URL:https://www.easa.europa.eu/en/domains/ drones-air-mobility/drones-air-mobility-landscape/urban-air-mobility-uam. accessed: 2026-05-14. Fein, D.J.,

    2026

  30. [30]

    Psychological Bulletin 76, 378–382

    Measuring nominal scale agreement among many raters. Psychological Bulletin 76, 378–382. doi:10.1037/h0031619. Fu, S., Avdelidis, N.P.,

    work page doi:10.1037/h0031619

  31. [31]

    Furman, B., Fabian, L., Ellis, S., Muller, P., Swenson, R.,

    doi:10.3390/s23198124. Furman, B., Fabian, L., Ellis, S., Muller, P., Swenson, R.,

    work page doi:10.3390/s23198124

  32. [32]

    Journal of Medical Internet Research 22, e16649

    Public Perception of Artificial Intelligence in Medical Care: Content Analysis of Social Media. Journal of Medical Internet Research 22, e16649. doi:10.2196/16649. Gao, Z., Yu, Y., Wei, Q., Topcu, U., Clarke, J.P.,

    work page doi:10.2196/16649

  33. [33]

    Transportation Research Part C: Emerging Technologies 165, 104740

    Noise-aware and equitable urban air traffic management: An optimization approach. Transportation Research Part C: Emerging Technologies 165, 104740. doi:10.1016/j.trc.2024.104740. Garcia, J.S., Truong, D., Hampton, S., Winter, S.R., Echevarne, R., Stolzer, A.J.,

    work page doi:10.1016/j.trc.2024.104740 2024

  34. [34]

    Transportation Research Interdisciplinary Perspectives 36, 101836

    Governance and effectiveness in aviation safety oversight: A structural model analysis. Transportation Research Interdisciplinary Perspectives 36, 101836. doi:10.1016/j.trip.2026.101836. Gers, F.A., Schmidhuber, J., Cummins, F.,

    work page doi:10.1016/j.trip.2026.101836 2026

  35. [35]

    Neural Computation 12, 2451–2471

    Learning to Forget: Continual Prediction with LSTM. Neural Computation 12, 2451–2471. doi:10.1162/089976600300015015. Giladi, R.,

    work page doi:10.1162/089976600300015015

  36. [36]

    Transportation Research Part D: Transport and Environment 87, 102527

    Real-time identification of aircraft sound events. Transportation Research Part D: Transport and Environment 87, 102527. doi:10.1016/j.trd.2020.102527. Go, A., Bhayani, R., Huang, L.,

    work page doi:10.1016/j.trd.2020.102527 2020

  37. [37]

    doi:10.3390/engproc2024080038. Hamilton,W.L.,Clark,K.,Leskovec,J.,Jurafsky,D.,2016.InducingDomain-SpecificSentimentLexiconsfromUnlabeledCorpora,in:Proceedings of the 2016 conference on empirical methods in natural language processing, Austin, TX, USA. pp. 595–605. Hasan, K.U.,

    work page doi:10.3390/engproc2024080038 2016

  38. [38]

    URL:https://thebatterytips.com/ battery-specifications/are-drones-battery-operated/

    Are drones battery operated? insights on energy sources and lifespan challenges. URL:https://thebatterytips.com/ battery-specifications/are-drones-battery-operated/. accessed: 2026-05-14. Hatzivassiloglou, V., McKeown, K.,

    2026

  39. [39]

    Predicting the Semantic Orientation of Adjectives, in: Proceedings of ACL, pp. 174–181. He, P., Liu, X., Gao, J., Chen, W., 2021a. DeBERTa: Decoding-enhanced BERT with Disentangled Attention. URL:https://arxiv.org/abs/ 2006.03654,arXiv:2006.03654. He, P., et al., 2021b. DeBERTa: Decoding-enhanced BERT with disentangled attention, in: ICLR. doi:10.48550/ar...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2006.03654 2006

  40. [40]

    Chinese Journal of Aeronautics 38, 103233

    A methodology for aircraft mission effectiveness evaluation and design space exploration based on virtual operational test. Chinese Journal of Aeronautics 38, 103233. doi:10.1016/j.cja.2024.09.009. Hu, M., Liu, B.,

    work page doi:10.1016/j.cja.2024.09.009 2024

  41. [41]

    Mining and summarizing customer reviews, in: Proceedings of the Tenth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Association for Computing Machinery, New York, NY, USA. p. 168–177. URL:https://doi.org/ 10.1145/1014052.1014073, doi:10.1145/1014052.1014073. Hugging Face, Accessed: Mar

    work page doi:10.1145/1014052.1014073

  42. [42]

    Hutto,C.,Gilbert,E.,2014a.Vader:Aparsimoniousrule-basedmodelforsentimentanalysisofsocialmediatext,in:Proceedingsoftheinternational AAAI conference on web and social media, pp

    Languages Supported on Hugging Face.https://huggingface.co/languages. Hutto,C.,Gilbert,E.,2014a.Vader:Aparsimoniousrule-basedmodelforsentimentanalysisofsocialmediatext,in:Proceedingsoftheinternational AAAI conference on web and social media, pp. 216–225. doi:10.1609/icwsm.v8i1.14550. Hutto, C., Gilbert, E., 2014b. VADER: A Parsimonious Rule-Based Model fo...

    work page doi:10.1609/icwsm.v8i1.14550

  43. [43]

    accessed: 2026-05-14

    URL:https://www.quantumrun.com/consulting/ twitter-users-statistics/. accessed: 2026-05-14. Jalandharachari, A., Bindhani, S., Kasodhan, P., Harish, D., Murthy, B.R., et al.,

    2026

  44. [44]

    Journal of the Air Transport Research Society , 100104doi:10.1016/j.jatrs.2026

    Integrating digital competency into aviation training: An auditable inputs–processes–outcomes framework. Journal of the Air Transport Research Society , 100104doi:10.1016/j.jatrs.2026. 100104. Jenkins, A.,

    work page doi:10.1016/j.jatrs.2026 2026

  45. [45]

    International Journal of Lifelong Education 40, 215–228

    Patterns of participation and non-participation in learning in mid-life and their determinants. International Journal of Lifelong Education 40, 215–228. doi:10.1080/02601370.2021.1937357. Jeong,D.,Hwang,S.,Kim,J.,Yu,H.,Park,E.,2023. Publicperspectiveonrenewableandotherenergyresources:Evidencefromsocialmediabig data and sentiment analysis. Energy Strategy ...

    work page doi:10.1080/02601370.2021.1937357 2021

  46. [46]

    doi:10.18653/v1/P17-1052

    Deep Pyramid Convolutional Neural Networks for Text Categorization, in: Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Vancouver, Canada. doi:10.18653/v1/P17-1052. Johnson, W., Silva, C.,

    work page doi:10.18653/v1/p17-1052

  47. [47]

    Kelemen,M.,Polishchuk,V.,Sabo,J.,2023

    Impact of Technology and Mission Variations on Aircraft Designed for Advanced Air Mobility, in: 13th Annual Electric VTOL Symposium, San Jose, CA, USA. Kelemen,M.,Polishchuk,V.,Sabo,J.,2023. AHybridModelforEvaluatingtheOutcomesofStudentPilotswithintheDidacticSystemofAviation Education. Aerospace

    2023

  48. [48]

    Kiesewetter, L., Shakib, K.H., Singh, P., Rahman, M., Khandelwal, B., Kumar, S., Shah, K.,

    doi:10.3390/aerospace10030281. Kiesewetter, L., Shakib, K.H., Singh, P., Rahman, M., Khandelwal, B., Kumar, S., Shah, K.,

    work page doi:10.3390/aerospace10030281

  49. [49]

    Progress in Aerospace Sciences 142, 100949

    A holistic review of the current state of research on aircraft design concepts and consideration for advanced air mobility applications. Progress in Aerospace Sciences 142, 100949. doi:10.1016/j.paerosci.2023.100949. Kim, Y., 2014a. Convolutional Neural Networks for Sentence Classification, in: EMNLP, Doha, Qatar. Kim,Y.,2014b. Convolutionalneuralnetworks...

    work page doi:10.1016/j.paerosci.2023.100949 2023

  50. [50]

    NASA Document ID: 20190000370

    Unmanned aircraft system traffic management (UTM) concept of operations, in: AIAA Aviation Forum and Exposition, Washington, DC, US. NASA Document ID: 20190000370. Kumar,A.,Abhishek,K.,ShafeeqBM,A.,2025.SentXFormer:atransformer-enhancedhybriddeeplearningframeworkforcross-domainsentiment analysis of customer reviews. Scientific Reports doi:10.1038/s41598-0...

    work page doi:10.1038/s41598-025-33526-1 2025

  51. [51]

    URL:https://arxiv.org/abs/1909.11942,arXiv:1909.11942

    ALBERT: A Lite BERT for Self-supervised Learning of Language Representations. URL:https://arxiv.org/abs/1909.11942,arXiv:1909.11942. LeCun, Y., Bengio, Y., Hinton, G.,

    Pith/arXiv arXiv 1909

  52. [52]

    LeCun, Y

    Deep learning. Nature 521, 436–444. doi:10.1038/nature14539. Lee,C.,Bae,B.,Lee,Y.L.,Pak,T.Y.,2023. Societalacceptanceofurbanairmobilitybasedonthetechnologyadoptionframework. Technological Forecasting and Social Change 196, 122807. doi:10.1016/j.techfore.2023.122807. Lee, D.D., Seung, H.S.,

    work page doi:10.1038/nature14539 2023

  53. [53]

    Nature 401, 788–791

    Learning the Parts of Objects by Non-negative Matrix Factorization. Nature 401, 788–791. doi:10.1038/44565. Li, Q.,

    work page doi:10.1038/44565

  54. [54]

    Research on Text Sentiment Classification and Recognition Based on Improved BiLSTM and TextCNN, in: 2024 2nd International Conference on Mechatronics, IoT and Industrial Informatics (ICMIII), IEEE. pp. 210–219. doi:10.1109/ICMIII62623.2024.00046. Li, W., Hu, X., Liu, Y.,

    work page doi:10.1109/icmiii62623.2024.00046 2024

  55. [55]

    TransparencyforSafety

    Joint sentiment/topic model for sentiment analysis, in: Proceedings of the 18th ACM Conference on Information and Knowledge Management, Association for Computing Machinery, New York, NY, USA. p. 375–384. URL:https://doi.org/10.1145/ 1645953.1646003, doi:10.1145/1645953.1646003. Lindhout,P.,Reniers,G.,2022. The“TransparencyforSafety”Triangle:DevelopingaSma...

    work page doi:10.1145/1645953.1646003 2022

  56. [56]

    Morgan & Claypool

    Sentiment Analysis and Opinion Mining. Morgan & Claypool. Liu,J.,Sun,S.,Tang,K.,Fan,X.,Lv,J.,Fu,Y.,Feng,X.,Zeng,L.,2025. IoT-BasedAirportNoisePerceptionandMonitoring:Multi-SourceData Fusion, Spatial Distribution Modeling, and Analysis. Sensors 25,

    2025

  57. [57]

    RoBERTa: A Robustly Optimized BERT Pretraining Approach

    doi:10.3390/s25082347. Liu,Y.,Ott,M.,Goyal,N.,Du,J.,Joshi,M.,Chen,D.,Levy,O.,Lewis,M.,Zettlemoyer,L.,Stoyanov,V.,2019. RoBERTa:ARobustlyOptimized BERT Pretraining Approach, in: arXiv preprint arXiv:1907.11692. doi:10.48550/arXiv.1907.11692. Liu, Z., Wang, Y., Kasai, J., Hajishirzi, H., Smith, N.A.,

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3390/s25082347 2019

  58. [58]

    Probing Across Time: What Does RoBERTa Know and When?, in: Findings of the Association for Computational Linguistics: EMNLP 2021, Punta Cana, Dominican Republic. pp. 820–842. Madigan, R., Louw, T., Wilbrink, M., Schieben, A., Merat, N.,

    2021

  59. [59]

    Transportation Research Part F: Traffic Psychology and Behaviour 50, 55–64

    What influences the decision to use automated public transport? using utaut to understand public acceptance of automated road transport systems. Transportation Research Part F: Traffic Psychology and Behaviour 50, 55–64. doi:10.1016/j.trf.2017.07.007. Mankins, J.C.,

    work page doi:10.1016/j.trf.2017.07.007 2017

  60. [60]

    Acta Astronautica 65, 1216–1223

    Technology readiness assessments: A retrospective. Acta Astronautica 65, 1216–1223. doi:10.1016/j.actaastro.2009. 03.058. Mao, Y., Liu, Q., Zhang, Y.,

    work page doi:10.1016/j.actaastro.2009 2009

  61. [61]

    Journal of King Saud University-Computer and Information Sciences 36, 102048

    Sentiment analysis methods, applications, and challenges: A systematic literature review. Journal of King Saud University-Computer and Information Sciences 36, 102048. doi:10.1016/j.jksuci.2024.102048. Marini,L.,Tonetti,M.S.,Nibali,L.,Rojas,M.A.,Aimetti,M.,Cairo,F.,Cavalcanti,R.,Crea,A.,Ferrarotti,F.,Graziani,F.,etal.,2021. Thestaging and grading system i...

    work page doi:10.1016/j.jksuci.2024.102048 2024

  62. [62]

    Journal of Infection and Public Health 14, 1505–1512

    Public sentiment analysis and topic modeling regarding covid-19 vaccines on the reddit social media platform: A call to action for strengthening vaccine confidence. Journal of Infection and Public Health 14, 1505–1512. doi:https://doi.org/10.1016/j.jiph.2021.08.010. Michel, A.H., Gettinger, D.,

    work page doi:10.1016/j.jiph.2021.08.010 2021

  63. [63]

    URL:https://dronecenter.bard.edu/2015-drone-year-in-review/

    2015 Drone Year in Review. URL:https://dronecenter.bard.edu/2015-drone-year-in-review/. accessed: 2026-04-02. Minaee,S.,Kalchbrenner,N.,Cambria,E.,Nikzad,N.,Chenaghlu,M.,Gao,J.,2021. DeepLearning–basedTextClassification:AComprehensive Review. ACM Computing Surveys

    2015

  64. [64]

    Moons, F., Vandervieren, E.,

    doi:10.1145/3439726. Moons, F., Vandervieren, E.,

    work page doi:10.1145/3439726

  65. [65]

    Mullen, T., Collier, N.,

    doi:10.3758/s13428-025-02746-8. Mullen, T., Collier, N.,

    work page doi:10.3758/s13428-025-02746-8

  66. [66]

    Accident Analysis & Prevention 159, 106228

    How can regulatory authorities improve safety in organizations by influencing safety culture? A conceptual model of the relationships and a discussion of implications. Accident Analysis & Prevention 159, 106228. doi:10.1016/j.aap. 2021.106228. First Author et al.:Preprint submitted to ElsevierPage 35 of 38 Short Title of the Article Nakshi, P., Palm, M., ...

    work page doi:10.1016/j.aap 2021

  67. [67]

    Transportation Research Record , 03611981251378483doi:10.1177/03611981251378483

    Using Questionnaires to Identify Travel Barriers: How (Not) to Ask Questions. Transportation Research Record , 03611981251378483doi:10.1177/03611981251378483. National Academies of Sciences, Engineering, and Medicine,

    work page doi:10.1177/03611981251378483

  68. [68]

    Built Environment Project and Asset Management 14, 278–295

    Significant risks that trigger cost overruns and delays in urban rail projects: a typical case study of Vietnam. Built Environment Project and Asset Management 14, 278–295. doi:10.1108/BEPAM-01-2023-0027. Nguyen, T.H., Shirai, K., Velcin, J.,

    work page doi:10.1108/bepam-01-2023-0027 2023

  69. [69]

    Expert Systems with Applications 42, 9603–9611

    Sentiment analysis on social media for stock movement prediction. Expert Systems with Applications 42, 9603–9611. doi:10.1016/j.eswa.2015.07.052. Nilsson,S.,2024. AdvancedAirMobility(AAM)isComingtoaTownNearYou...AreYouReady?,in:UASSummit&Expo,GrandForks,North Dakota, USA. Noori, B.,

    work page doi:10.1016/j.eswa.2015.07.052 2015

  70. [70]

    Applied Artificial Intelligence 35, 567–588

    Classification of Customer Reviews Using Machine Learning Algorithms. Applied Artificial Intelligence 35, 567–588. doi:10.1080/08839514.2021.1922843. Nordhoff,S.,deWinter,J.,Kyriakidis,M.,vanArem,B.,Happee,R.,2019. Amused,accepted,andused?attitudesandemotionstowardsautomated vehicles, their relationships, and predictive value for usage intention. Transpor...

    work page doi:10.1080/08839514.2021.1922843 2021

  71. [71]

    Paris: Organisation for Economic Co-operation and Development doi:10.1787/9789264311756-en

    Getting Skills Right: Future-Ready Adult Learning Systems. Paris: Organisation for Economic Co-operation and Development doi:10.1787/9789264311756-en. OpenAI,

    work page doi:10.1787/9789264311756-en

  72. [72]

    Accessed: 2026-03-

    OpenAI API Documentation.https://platform.openai.com/docs/api-reference/introduction. Accessed: 2026-03-

    2026

  73. [73]

    AMachineLearning-BasedUsabilityEvaluationMethodforeLearningSystems

    Oztekin,A.,Zaim,S.,Delen,D.,2013. AMachineLearning-BasedUsabilityEvaluationMethodforeLearningSystems. DecisionSupportSystems 56, 63–73. doi:10.1016/j.dss.2013.05.003. Pang, B., Lee, L., Vaithyanathan, S.,

    work page doi:10.1016/j.dss.2013.05.003 2013

  74. [74]

    Thumbs up? sentiment classification using machine learning techniques, in: Proceedings of the 2002 Conference on Empirical Methods in Natural Language Processing (EMNLP 2002), Association for Computational Linguistics. pp. 79–86. doi:10.3115/1118693.1118704. Park, E., Lee, K., Han, T., Nam, H.S.,

    work page doi:10.3115/1118693.1118704 2002

  75. [75]

    Pongsakornsathien,N.,Safwat,N.E.D.,Xie,Y.,Gardi,A.,Sabatini,R.,2025

    doi:10.3390/jpm12010020. Pongsakornsathien,N.,Safwat,N.E.D.,Xie,Y.,Gardi,A.,Sabatini,R.,2025. Advancesinlow-altitudeairspacemanagementforuncrewedaircraft and advanced air mobility. Progress in Aerospace Sciences 154, 101085. doi:10.1016/j.paerosci.2025.101085. Pons-Prats, J., Živojinović, T., Kuljanin, J.,

    work page doi:10.3390/jpm12010020 2025

  76. [76]

    TransportationResearchPartE:LogisticsandTransportationReview166,102868

    On the understanding of the current status of urban air mobility development and its future prospects:Commutinginaflyingvehicleasanewparadigm. TransportationResearchPartE:LogisticsandTransportationReview166,102868. doi:10.1016/j.tre.2022.102868. Poria,S.,Cambria,E.,Gelbukh,A.,2017. Context-DependentSentimentAnalysisinUser-GeneratedVideos. ACLdoi:10.18653/...

    work page doi:10.1016/j.tre.2022.102868 2022

  77. [77]

    Accessed 2026-03-03

    PRAW Documentation.https://praw.readthedocs.io. Accessed 2026-03-03. PublicCEO,

    2026

  78. [78]

    URL:https://www.publicceo.com/2016/03/ drones-a-growing-hazard-in-the-absence-of-tighter-regulations/

    Drones: A growing hazard in the absence of tighter regulations. URL:https://www.publicceo.com/2016/03/ drones-a-growing-hazard-in-the-absence-of-tighter-regulations/. Qiu, G., Liu, B., Bu, J., Chen, C.,

    2016

  79. [79]

    Computational Linguistics 37, 9–27

    Opinion Word Expansion and Target Extraction through Double Propagation. Computational Linguistics 37, 9–27. doi:10.1162/coli_a_00034. Quora Inc.,

    work page doi:10.1162/coli_a_00034

  80. [80]

    Raza,W.,Renkhoff,J.,Ogirimah,O.,Bawa,G.K.,Stansbury,R.S.,2025

    Quora for Business.https://business.quora.com/. Raza,W.,Renkhoff,J.,Ogirimah,O.,Bawa,G.K.,Stansbury,R.S.,2025. AdvancedAirMobility:Innovations,Applications,Challenges,andFuture Potential. Journal of Air Transportation 33, 169–187. doi:10.2514/1.D0440. Reddit Inc.,

    work page doi:10.2514/1.d0440 2025

Showing first 80 references.