t-SNE Manifold Learning Based Visualization: A Human Activity Recognition Approach

  • Ramesh DharavathEmail author
  • G. MadhukarRao
  • Himanshu Khurana
  • Damodar Reddy Edla
Conference paper
Part of the Lecture Notes on Data Engineering and Communications Technologies book series (LNDECT, volume 37)


Visualization of data which is enormously generated at present has become one of the major steps in machine learning before applying the classifier on related domain problems. This paper inculcates the visualization and recognition of human activity from n-dimensional space to low-dimensional space by implementing an algorithm and sometimes best termed as a tool called as t-SNE (t-distributed stochastic neighborhood embedding), which is a non-linear algorithm and has been inspired by SNE (stochastic neighbor embedding). This algorithm is applied to build the relationship between various actions done by human, where class label information has been introduced to well represent the actions from similar action class. This comes under a machine learning methodology called manifold learning, where the class label information helps to classify the action done by human. This learning technique uses probabilistic embedding of neighbors from n-dimensional space with low-dimension to find a mapping of data points from one distribution to another distribution. Experimental results and their comparison with other visualization tools like PCA shows the effectiveness of t-SNE algorithm with human activity recognition and MNIST dataset. The results produced by t-SNE algorithm in terms of visualization are always better than many other algorithms.


Dimensionality reduction Manifold learning Stochastic neighbor embedding Visualization 



This work is partially supported by the Indian Institute of Technology (ISM), Dhanbad that comes under the administrative and financial control of the Ministry of Human Resource Development (MHRD), Government of India. The authors express their gratitude towards the Department of Computer Science and Engineering at IIT (ISM) for providing all the necessary support to carry out the research work.


  1. 1.
    N. Ravi, N. Dandekar, P. Mysore, M.L. Littman, Activity recognition from accelerometer data. In Aaai, vol. 5, no. 2005 (2005, July), pp. 1541–1546Google Scholar
  2. 2.
    X. Liu, T. You, X. Ma, H. Kuang, An optimization model for human activity recognition inspired by information on human-object interaction. In 2018 10th International Conference on Measuring Technology and Mechatronics Automation (ICMTMA) (IEEE, 2018, February), pp. 519–523Google Scholar
  3. 3.
    Y. Guan, T. Plötz, Ensembles of deep lstm learners for activity recognition using wearables. Proc. ACM Interact Mob. Wearable Ubiquitous Technol. 1(2), 11 (2017)CrossRefGoogle Scholar
  4. 4.
    E. Kantoch, D. Grochala, M. Kajor, Bio-inspired topology of wearable sensor fusion for telemedical application. In International Conference on Artificial Intelligence and Soft Computing (Springer, Cham, 2017, June), pp. 658–667Google Scholar
  5. 5.
    L.V.D. Maaten, G. Hinton, Visualizing data using t-SNE. J. Mach. Learn. Res. 9(Nov), 2579–2605 (2008)Google Scholar
  6. 6.
    Z. Zhongwen, G. Huanghuang, Visualization study of high-dimensional data classification based on PCA-SVM. In 2017 IEEE Second International Conference on Data Science in Cyberspace (DSC) (IEEE, 2017, June), pp. 346–349)Google Scholar
  7. 7.
    R. Bhardwaj, K. Dang, S.C. Gupta, S. Kumar, Review on human activity recognition using soft computing. In Soft Computing: Theories and Applications (Springer, Singapore, 2018), pp. 783–790Google Scholar
  8. 8.
    J. Jo, J. Seo, J.D. Fekete, A progressive kd tree for approximate k-nearest neighbors. In Workshop on Data Systems for Interactive Analysis (DSIA) (2017, October)Google Scholar
  9. 9.
    E. Krivov, M. Belyaev, Dimensionality reduction with isomap algorithm for EEG covariance matrices. In 2016 4th International Winter Conference on Brain-Computer Interface (BCI) (IEEE, 2016, February), pp. 1–4Google Scholar
  10. 10.
    J. Cheng, H. Liu, F. Wang, H. Li, C. Zhu, Silhouette analysis for human action recognition based on supervised temporal t-SNE and incremental learning. IEEE Trans. Image Process 24(10), 3203–3217 (2015)MathSciNetCrossRefGoogle Scholar
  11. 11.
    N. Rogovschi, J. Kitazono, N. Grozavu, T. Omori, S. Ozawa, t-Distributed stochastic neighbor embedding spectral clustering. In 2017 International Joint Conference on Neural Networks (IJCNN) (IEEE, 2017, May) (pp. 1628–1632)Google Scholar
  12. 12.
    E. Kantoch, Human activity recognition for physical rehabilitation using wearable sensors fusion and artificial neural networks. In Computing in Cardiology (CinC), 2017 (IEEE, 2017, September), pp. 1–4Google Scholar
  13. 13.
    Y. Hou, H. Zhao, Handwritten digit recognition based on depth neural network. In 2017 International Conference on Intelligent Informatics and Biomedical Sciences (ICIIBMS) (IEEE, 2017, November), pp. 35–38Google Scholar
  14. 14.
    T. Dungkaew, J. Suksawatchon, U. Suksawatchon, Impersonal smartphone-based activity recognition using the accelerometer sensory data. In 2017 2nd International Conference on Information Technology (INCIT) (IEEE, 2017, November), pp. 1–6Google Scholar
  15. 15.
    K. Nakano, B. Chakraborty, Effect of dynamic feature for human activity recognition using smartphone sensors. In 2017 IEEE 8th International Conference on Awareness Science and Technology (iCAST) (IEEE, 2017, November), pp. 539–543Google Scholar
  16. 16.
    S.T. Roweis, L.K. Saul, Nonlinear dimensionality reduction by locally linear embedding. Science 290(5500), 2323–2326 (2000)CrossRefGoogle Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2020

Authors and Affiliations

  1. 1.Department of Computer Science and EngineeringIndian Institute of Technology (Indian School of Mines)DhanbadIndia
  2. 2.Department of Computer Science and EngineeringNational Institute of TechnologyGoaIndia

Personalised recommendations