Abstract
The introduction of interactive variants of Non-negative Matrix Factorization and Latent Dirichlet Allocation algorithms has made it possible to develop systems that support human in the loop interaction and refinement of topic models. This paper presents design guidelines for developing software that is able to interactively support researchers and decision makers in the discovery and interpretation of topics found in both small and large domain-specific document collections. The guidelines emerged from the analysis of two between-subjects experiments, which involved comparing algorithmically derived topics with the topics produced by research analysts and evaluating interactive topic modeling algorithms. A key finding from both experiments was that a holistic approach that includes additional algorithms needs to be taken when designing human in the loop systems to support research and decision making. The paper also outlines key areas where additional research is required in order to comply with the proposed design guidelines. Much of the proposed functionality is not currently included in topic model exploration software.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Andrzejewski, D., Zhu, X., Craven, M.: Incorporating domain knowledge into topic modeling via dirichlet forest priors. In: Proceedings of the 26th Annual International Conference on Machine Learning, pp. 25–32. ACM (2009)
Bakharia, A.: Interactive content analysis: evaluating interactive variants of non-negative Matrix Factorisation and Latent Dirichlet Allocation as qualitative content analysis aids. Ph.D. thesis, Queensland University of Technology (2014)
Bakharia, A., Bruza, P., Watters, J., Narayan, B., Sitbon, L.: Interactive topic modeling for aiding qualitative content analysis. In: Proceedings of the 2016 ACM on Conference on Human Information Interaction and Retrieval, pp. 213–222. ACM (2016)
Blei, D.M., Ng, A.Y., Jordan, M.I.: Latent dirichlet allocation. J. Mach. Learn. Res. 3, 993–1022 (2003)
Boutsidis, C., Gallopoulos, E.: SVD based initialization: a head start for nonnegative matrix factorization. Pattern Recogn. 41(4), 1350–1362 (2008)
Lee, T.Y., Smith, A., Seppi, K., Elmqvist, N., Boyd-Graber, J., Findlater, L.: The human touch: how non-expert users perceive, interpret, and fix topic models. Int. J. Hum Comput Stud. 105, 28–42 (2017)
Mei, Q., Shen, X., Zhai, C.: Automatic labeling of multinomial topic models. In: Proceedings of the 13th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 490–499. ACM (2007)
Mikolov, T., Sutskever, I., Chen, K., Corrado, G.S., Dean, J.: Distributed representations of words and phrases and their compositionality. In: Advances in Neural Information Processing Systems, pp. 3111–3119 (2013)
Mimno, D., Wallach, H.M., Talley, E., Leenders, M., McCallum, A.: Optimizing semantic coherence in topic models. In: Proceedings of the Conference on Empirical Methods in Natural Language Processing, pp. 262–272. Association for Computational Linguistics (2011)
Newman, D., Noh, Y., Talley, E., Karimi, S., Baldwin, T.: Evaluating topic models for digital libraries. In: Proceedings of the 10th Annual Joint Conference on Digital Libraries, pp. 215–224. ACM (2010)
O’Callaghan, D., Greene, D., Carthy, J., Cunningham, P.: An analysis of the coherence of descriptors in topic modeling. Expert Syst. Appl. 42(13), 5645–5657 (2015)
Seung, D., Lee, L.: Algorithms for non-negative matrix factorization. Adv. Neural Inf. Proc. Syst. 13, 556–562 (2001)
Sievert, C., Shirley, K.E.: LDAvis: A method for visualizing and interpreting topics (2014)
Wang, F., Li, T., Zhang, C.: Semi-supervised clustering via matrix factorization. In: SDM, pp. 1–12 (2008)
Acknowledgements
The experiments described within this paper were conducted as part of my doctorate degree at Queensland University of Technology. I would like to thank and acknowledge my supervisors Peter Bruza, Jim Watters, Bhuva Narayan and Laurianne Sitbon.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer International Publishing AG, part of Springer Nature
About this paper
Cite this paper
Bakharia, A. (2019). Designing Interactive Topic Discovery Systems for Research and Decision Making. In: Czarnowski, I., Howlett, R., Jain, L., Vlacic, L. (eds) Intelligent Decision Technologies 2018. KES-IDT 2018 2018. Smart Innovation, Systems and Technologies, vol 97. Springer, Cham. https://doi.org/10.1007/978-3-319-92028-3_1
Download citation
DOI: https://doi.org/10.1007/978-3-319-92028-3_1
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-92027-6
Online ISBN: 978-3-319-92028-3
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)