Advertisement

A Machine Learning Approach to Predict Diabetes Using Short Recorded Photoplethysmography and Physiological Characteristics

  • Chirath HettiarachchiEmail author
  • Charith Chitraranjan
Conference paper
  • 803 Downloads
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11526)

Abstract

Diabetes is a global epidemic, which leads to severe complications such as heart disease, limb amputations and blindness, mainly occurring due to the inability of early detection. Photoplethysmography (PPG) signals have been used as a non-invasive approach to predict diabetes. However, current methods use long, continuous signals collected in a clinical setting. This study focuses on predicting Type 2 Diabetes from short (~2.1s) PPG signals extracted from smart devices, and readily available physiological data such as age, gender, weight and height. Since this type of PPG signals can be easily extracted using mobile phones or smart wearable technology, the user can get an initial prediction without entering a medical facility. Through the analysis of morphological features related to the PPG waveform and its derivatives, we identify features related to Type 2 Diabetes and establish the feasibility of predicting Type 2 Diabetes from short PPG signals. We cross validated several classification models based on the selected set of features to predict Type 2 Diabetes, where Linear Discriminant Analysis (LDA) achieved the highest area under the ROC curve of 79%. The successful practical implementation of the proposed system would enable people to screen themselves conveniently using their smart devices to identify the potential risk of Type 2 Diabetes and thus avoid austere complications of late detection.

Keywords

Machine Learning Diabetes Type II Photoplethysmography Feature selection 

Notes

Acknowledgement

This research was partially funded by the Senate Research Council grant number SRC/LT/2019/33 of University of Moratuwa, Sri Lanka.

References

  1. 1.
    Elgendi, M.: On the analysis of fingertip photoplethysmogram signals. Curr. Cardiol. Rev. 8(1), 14–25 (2012)CrossRefGoogle Scholar
  2. 2.
    Allen, J.: Photoplethysmography and its application in clinical physiological measurement. Physiol. Meas. 28(3), R1 (2007)CrossRefGoogle Scholar
  3. 3.
    Reddy, V.R., et al.: DMSense: a non-invasive diabetes mellitus classification system using photoplethysmogram signal. In: IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops). IEEE (2017)Google Scholar
  4. 4.
    Ballinger, B., et al.: DeepHeart: semi-supervised sequence learning for cardiovascular risk prediction. In: Thirty-Second AAAI Conference on Artificial Intelligence (2018)Google Scholar
  5. 5.
    Moreno, E.M., et al.: Type 2 diabetes screening test by means of a pulse oximeter. IEEE Trans. Biomed. Eng. 64(2), 341–351 (2017)CrossRefGoogle Scholar
  6. 6.
    Liang, Y., et al.: A new, short-recorded photoplethysmogram dataset for blood pressure monitoring in China. Sci. Data 5, 180020 (2018)CrossRefGoogle Scholar
  7. 7.
    Elgendi, M.: Optimal signal quality index for photoplethysmogram signals. Bioengineering 3(4), 21 (2016)CrossRefGoogle Scholar
  8. 8.
    Liang, Y., et al.: An optimal filter for short photoplethysmogram signals. Sci. Data 5, 180076 (2018)CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Department of Computer Science and EngineeringUniversity of MoratuwaMoratuwaSri Lanka

Personalised recommendations