How to use t-SNE effectively
Open accessAbstract
The purpose of this thesis is to evaluate unsupervised clustering strategies for data associated with bulk carrier trucks, in the detection of failure patterns in thermal systems.The study of these techniques is important in the field of data-based predictive maintenance with the implementation of machine learning algorithms that allow proper planning of maintenance schedules in freight transport companies.For the development of the thesis, the telemetry devices of the bulk tractor trucks of a Colombian cargo transport company are used as a source of information, which report data in real time of the measurement of variables such as speed, temperatures, state of operation of the vehicle, among others for the year 2020.The history of workshop entries of the fleet of 116 tractor-trailers is also used, where workshop entries for the intervention of thermal systems are analyzed.These data are the input for the evaluation of the grouping strategies proposed in this work.The results start from obtaining the data, preparing them and descriptive analysis to implement dimensionality reduction techniques in the information and subsequently evaluate the behavior of grouping algorithms for the detection of failure patterns that are related to damage in thermal systems.With the development of this work, there is a potential for savings in corrective costs of the fleet in the workshop that points to an adequate management of the fleet in pay-per-use models, leveraging the availability of vehicles in transport operations.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Unlike t-SNE or UMAP, PCA offers a deterministic, global projection that avoids local instability and sensitivity to initialization.
“Unlike manifold learning techniques such as t-SNE or UMAP, PCA offers a deterministic, global projection that avoids issues of local instability and sensitivity to initialization (Wattenberg et al., 2016).”
From Can "consciousness" be observed from large language model (LLM) internal states? Dissecting LLM representations obtained from Theory of Mind test with Integrated Information Theory and Span Representation analysis (2025) · cites this paper as Wattenberg et al., 2016
A difficulty with t-SNE is the choice of perplexity parameter, as graphs may show different numbers of clusters depending on its value.
“A further difficulty is the choice of its perplexity parameter, as t‐SNE graphs may show strongly different numbers of clusters depending on its value (Wattenberg et al, 2016).”
From Current best practices in single-cell RNA-seq analysis: a tutorial (2019) · cites this paper as Wattenberg et al, 2016
Cites
(2008) · Journal of Machine Learning Research · Cited by 33,051 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Dimensionality reduction for visualizing single-cell data using UMAP
Becht, E., McInnes, L., Healy, J., Dutertre, C.-A., Kwok, I. W. H., Ng, L. G., Ginhoux, F., & Newell, E. W. (2019) · Nature Biotechnology · doi:10.1038/nbt.4314 · Cited by 5,358 · see citing works
Current best practices in single-cell RNA-seq analysis: a tutorial
Luecken, M. D., & Theis, F. J. (2019) · Molecular Systems Biology · doi:10.15252/msb.20188746 · Cited by 2,231 · see citing works
Interpretable machine learning: Fundamental principles and 10 grand challenges
Rudin, C., Chen, C., Chen, Z., Huang, H., Semenova, L., & Zhong, C. (2022) · Statistics Surveys · doi:10.1214/21-ss133 · Cited by 958 · see citing works
Visual Analytics in Deep Learning: An Interrogative Survey for the Next Frontiers
Hohman, F., Kahng, M., Pienta, R., & Chau, D. H. (2019) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2018.2843369 · Cited by 582 · see citing works
Interpretable dimensionality reduction of single cell transcriptome data with deep generative models
Ding, J., Condon, A., & Shah, S. P. (2018) · Nature Communications · doi:10.1038/s41467-018-04368-5 · Cited by 337 · see citing works
Borys, K., Schmitt, Y. A., Nauta, M., Seifert, C., Krämer, N., Friedrich, C. M., & Nensa, F. (2023) · European Journal of Radiology · doi:10.1016/j.ejrad.2023.110786 · Cited by 233 · see citing works
COVID-19 diagnosis by routine blood tests using machine learning
Kukar, M., Gunčar, G., Vovko, T., Podnar, S., Černelč, P., Brvar, M., Zalaznik, M., Notar, M., Moškon, S., & Notar, M. (2021) · Scientific Reports · doi:10.1038/s41598-021-90265-9 · Cited by 179 · see citing works
Global birdsong embeddings enable superior transfer learning for bioacoustic classification
Ghani, B., Denton, T., Kahl, S., & Klinck, H. (2023) · Scientific Reports · doi:10.1038/s41598-023-49989-z · Cited by 167 · see citing works
The landscape of biomedical research
González-Márquez, R., Schmidt, L., Schmidt, B. M., Berens, P., & Kobak, D. (2024) · Patterns · doi:10.1016/j.patter.2024.100968 · Cited by 89 · see citing works
Single-Cell Computational Strategies for Lineage Reconstruction in Tissue Systems
Herring, C. A., Chen, B., McKinley, E. T., & Lau, K. S. (2018) · Cellular and Molecular Gastroenterology and Hepatology · doi:10.1016/j.jcmgh.2018.01.023 · Cited by 35 · see citing works
Longitudinal cardio-respiratory fitness prediction through wearables in free-living environments
Spathis, D., Perez-Pozuelo, I., Gonzales, T. I., Wu, Y., Brage, S., Wareham, N., & Mascolo, C. (2022) · npj Digital Medicine · doi:10.1038/s41746-022-00719-1 · Cited by 30 · see citing works
Reynolds, K. E., Huang, E., Sabbineni, M., Wiseman, E., Murtaza, N., Ahuja, D., Napier, M., Murphy, K. M., Singh, K. K., & Scott, A. L. (2024) · Molecular Neurobiology · doi:10.1007/s12035-024-04181-w · Cited by 6 · see citing works
Li, J. (2025) · Natural Language Processing Journal · doi:10.1016/j.nlp.2025.100163 · Cited by 3 · see citing works
“Unlike manifold learning techniques such as t-SNE or UMAP, PCA offers a deterministic, global projection that avoids issues of local instability and sensitivity to initialization (Wattenberg et al., 2016).”
Improving Environment Novelty Quantification for Effective Unsupervised Environment Design
Teoh, J., Li, W., & Varakantham, P. (2025) · arXiv
Improving Environment Novelty Quantification for Effective Unsupervised Environment Design
Teoh, J., Li, W., & Varakantham, P. (2024) · Advances in Neural Information Processing Systems 37 · doi:10.52202/079017-4300
- Authors
- Wattenberg, M., Viégas, F., & Johnson, I.
- Affiliations
- Google (United States) (US)
- Google Cloud
- Venue
- Distill
- Publisher
- Distill Working Group
- Published
- 2016-10-13
- Volume
- 1
- Issue
- 10
- ISSN
- 2476-0757 (ISSN-L)
- OpenAlex
- W2542768043
- Cited by
- 854
- References
- 1
- Citations by year
- 2026: 62 · 2025: 93 · 2024: 101 · 2023: 130 · 2022: 144 · 2021: 98 · 2020: 84 · 2019: 74 · 2018: 46 · 2017: 20 · 2016: 2
- FWCI
- 41.10
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Distill (publishedVersion, cc-by)
- Topics
- Physical Sciences › Engineering › Automotive Engineering › Technical Engine Diagnostics and Monitoring
- Physical Sciences › Engineering › Control and Systems Engineering › Fault Detection and Control Systems
- Physical Sciences › Engineering › Control and Systems Engineering › Machine Fault Diagnosis Techniques
- Keywords
- t-SNE, unsupervised clustering, tractor-trailers, predictive maintenance, dimensionality reduction, pay-per-use
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
