Elamur Research⌕
☰
home / papers / 2026 / Tracking the Evolution of Foresight Signals in News Data: The Case of the European Electric Vehicle Market

Tracking the Evolution of Foresight Signals in News Data: The Case of the European Electric Vehicle Market

Paper · 2026 · Karine Navasartian

Summary

Work on NLP for strategic foresight has mostly treated weak signals as static, one-time detections. It has paid little attention to how signals develop and gain prominence over time. This student workshop paper tests whether following the temporal paths of signals helps judge their strategic value. The author built a multilingual news corpus on the European electric vehicle market from LexisNexis, covering five languages. Signals for each time step were extracted with BERTopic and linked across periods with a single-pass clustering algorithm. The resulting trajectories were then examined with burst detection. Three real cases served as validation: fears of Chinese espionage, job effects of the combustion-engine ban, and geopolitical dependence caused by critical-mineral shortages. The analysis finds three dominant trajectory types. Most signals appear only once, but some recur sparsely over long horizons. Burst metrics tend to flag signals that have already consolidated rather than giving early warning. The paper concludes that foresight practice needs dynamic, longitudinal monitoring and new metrics that can catch signals earlier.

Key points

  • Corpus: 56,994 European news articles (October 2018 – September 2025) retrieved through the LexisNexis API in English, German, Spanish, French and Italian.
    • Industry and source-rank filters were applied.
    • Duplicates were removed by title and URL.
    • Articles in the shortest and longest 2.5% of the length distribution were dropped.
  • Pipeline: BERTopic topic extraction using the ibm-granite/granite-embedding-107m-multilingual embedding model, followed by single-pass clustering to connect topics over time and burst detection (burstiness score > 0) on the trajectories.
  • Trajectory patterns: Most signals do not recur, which rules out trajectory analysis for them. Others follow irregular, gap-filled, bursty recurrence and often resurface years later as more established discourse.
  • Burst detection: In cases such as lithium supply risks and Chinese industrial espionage, bursts confirmed signal maturation instead of anticipating it.
  • Stated limitations: no external ground truth, a single mature domain, reliance on one temporal metric, and parameter choices tuned to this use case.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓University of BremenBremen · Germany6 works · 38,395 citations

 Fields, topics and concepts

📚 Shelves 8The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksSocial Sciences · 1,772 worksInformation Systems · 1,437 worksGeneral Social Sciences · 227 worksBusiness, Management and Accounting · 32 worksStrategy and Management · 4 works

Cite

Cite as Karine Navasartian. 2026. Tracking the Evolution of Foresight Signals in News Data: The Case of the European Electric Vehicle Market. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 4: Student Research Workshop).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-srw.28

@inproceedings{navasartian2026tracking,
  title = {{Tracking the Evolution of Foresight Signals in News Data: The Case of the European Electric Vehicle Market}},
  author = {Karine Navasartian},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 4: Student Research Workshop)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-srw.28},
  url = {https://doi.org/10.18653/v1/2026.acl-srw.28}
}

References

27 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Philipp Adämmer and Rainer A Schüssler. 2020. Forecasting the equity premium: mind the news! Review of Finance, 24(6):1313–1355. find
  2. H Igor Ansoff. 1975. Managing strategic surprise by response to weak signals. California management review, 18(2):21–33. find
  3. Claus Doll and Martin Wietschel. 2022. Aus für verbrennungsmotoren bis 2035: Statement zum beschluss des eu-parlaments zum ausstieg aus dem verbrennungsmotor für neue pkw und leichte nutz-fahrzeuge bis 2035. find
  4. Ashkan Ebadi, Alain Auger, and Yvan Gauthier. 2024a. Harnessing machine learning and stock market techniques for signal detection in underwater sensing technologies. In 2024 International Conference on Machine Learning and Applications (ICMLA), pages 1747–1752. IEEE. find
  5. Ashkan Ebadi, Alain Auger, and Yvan Gauthier. 2024b. Wisdom: An ai-powered framework for emerging research detection using weak signal analysis and advanced topic modeling. arXiv preprint arXiv:2409.15340. — [arXiv:2409.15340](https://arxiv.org/abs/2409.15340) find
  6. Roman Egger and Joanne Yu. 2022. A topic modeling comparison between lda, nmf, top2vec, and bertopic to demystify twitter posts. Frontiers in sociology, 7:886498. find
  7. Manal El Akrouchi, Houda Benbrahim, and Ismail Kassou. 2021. End-to-end lda-based automatic weak signal detection in web news. Knowledge-Based Systems, 212:106650. find
  8. Amber Geurts, Ralph Gutknecht, Philine Warnke, Arjen Goetheer, Elna Schirrmeister, Babette Bakker, and Svetlana Meissner. 2022. New perspectives for data-supported foresight: The hybrid ai-expert approach. Futures & Foresight Science, 4(1):e99. find
  9. Maarten Grootendorst. 2022. Bertopic: Neural topic modeling with a class-based tf-idf procedure. arXiv preprint arXiv:2203.05794.
  10. Benjamin Harris and Jennifer Brooker. 2025. Environmental scanning: A look to the future. New Directions for Evaluation, 2025(185-186):33–41. find
  11. Elina Hiltunen. 2008. The future sign and its three dimensions. Futures, 40(3):247–260. find
  12. Eun-Kyeong Kim and Hang-Hyun Jo. 2016. Measuring burstiness for finite event sequences. Physical Review E, 94(3):032311. find
  13. Julia Beißwenger. 2012. Rohstoffmangel könnte den vormarsch der elektroautos bremsen. Deutschlandfunk. find
  14. Allaa Boutaleb, Jérôme Picault, and Guillaume Grosjean. 2024. Bertrend: neural topic modeling for emerging trends detection. arXiv preprint arXiv:2411.05930. find
  15. Patrick Brandtner and Marius Mates. 2021. Artificial intelligence in strategic foresight–current practices and future application potentials: current practices and future application potentials. In Proceedings of the 2021 12th International Conference on E-business, Management and Economics, pages 7… find
  16. Jon Kleinberg. 2002. Bursty and hierarchical structure in streams. In Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining, pages 91–101. find
  17. Kevin Kohler. 2021. Strategic foresight: Knowledge, tools, and methods for the future. CSS Risk and Resilience Reports. find
  18. MTEB Leaderboard. 2025. https://huggingface.co/spaces/mteb/leaderboard. Accessed: 2025-03-12. ↗
  19. LexisNexis Developer Portal. 2025. https://dev.lexisnexis.com/home. Accessed: 2025-04-02. ↗
  20. Christian Mühlroth and Michael Grottke. 2018. A systematic literature review of mining weak signals and trends for corporate foresight. Journal of Business Economics, 88(5):643–687. find
  21. Christian Mühlroth, Laura Kölbl, and Michael Grottke. 2023. Innovation signals: leveraging machine learning to separate noise from news. Scientometrics, 128(5):2649–2676. find
  22. Vinicius Muraro and Sergio Salles-Filho. 2024. Big data, machine learning and uncertainty in foresight studies. foresight, 26(3):436–452. find
  23. Judicael Poumay and Ashwin Ittoo. 2022. Using meaning instead of words to track topics. In International Conference on Applications of Natural Language to Information Systems, pages 461–468. Springer. find
  24. Pauline Rousseau, Daniel Camara, and Dimitris Kotzinos. 2021. Weak signal detection and identification in large data sets: a review of methods and applications. find
  25. Hyunsang Son and Young Eun Park. 2025. Agenda-setting effects for covid-19 vaccination: Insights from 10 million textual data from social media and news articles using bertopic. International Journal of Information Management, 83:102907. find
  26. Christoph Vaalma, Daniel Buchholz, Marcel Weil, and Stefano Passerini. 2018. A cost and resource analysis of sodium-ion batteries. Nature reviews materials, 3(4):1–11. find
  27. Yongqi Wang and Yujie Xie. 2025. Legitimation contestation of the “chinese car threat” narrative in the digitized public sphere: a critical discourse analysis. Humanities and Social Sciences Communications, 12(1):1923. sourceRank:(1 2 3) AND industry:Automotive AND sourceRegion:Europe AND language:(… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/navasartian-2026-tracking-evolution-foresight-signals-news/