A stacked CNN and random forest ensemble architecture for complex nursing activity recognition and nurse identification
2024

Recognizing Nursing Activities with AI

Sample size: 8 publication 10 minutes Evidence: high

Author Information

Author(s): Rahman Arafat, Nahid Nazmun, Schuller Björn, Ahad Md Atiqur Rahman

Primary Institution: Kyushu Institute of Technology, Kitakyushu, Japan

Hypothesis

Can a stacked CNN and random forest ensemble accurately recognize complex nursing activities and identify nurses?

Conclusion

The proposed algorithm achieved high accuracy in recognizing nursing activities and identifying users simultaneously.

Supporting Evidence

  • The algorithm achieved 70.6% accuracy for activity recognition on the CARECOM dataset.
  • The highest accuracy for user identification was 92.7% on the Heiseikai dataset.
  • The study introduced a novel feature called mean min max sum to improve recognition accuracy.
  • The ensemble model outperformed traditional CNNs in recognizing complex nursing activities.

Takeaway

This study created a smart system that can tell what nurses are doing and who they are, using data from their smartphones.

Methodology

The study used a two-step feature extraction method and an ensemble of a random forest classifier and a stacked CNN model to analyze data from two benchmark nurse care activity datasets.

Potential Biases

Potential biases may arise from the limited number of participants and the specific contexts in which data were collected.

Limitations

The study faced challenges with class imbalance and the complexity of nursing activities, which may affect the generalizability of the results.

Participant Demographics

All participants were Japanese nurses.

Digital Object Identifier (DOI)

10.1038/s41598-024-81228-x

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication