A naïve Gaussian Bayes classifier for detection of mental activity in gait signature

Abstract

A probabilistic modelling is presented to detect mental activity from gait signature recorded from healthy subjects. The proposed scheme is based on principal component analysis with reduced feature dimension followed by a naïve Gaussian Bayes classifier. The leave-one-out cross-validation shows the detection accuracy of 94% with specificity and sensitivity of 96% and 98.3%, respectively. The research has a potential application in the prevention of elderly risk falls, lie detection and rehabilitation among Parkinson's patients.

Keywords:

• probabilistic modelling
• gait signature
• mental activity
• principal component analysis
• leave-one-out cross-validation

Acknowledgements

The authors highly acknowledge the volunteers who participated in this study. We also sincerely thank the scientists and technical staffs at Ergonomics Lab, Defense Institute of Physiological and Allied Sciences (DIPAS), Timarpur, New Delhi.

Notes

1. Email: [email protected]

2. Email: [email protected]

Additional information

Notes on contributors

A. Mishra

1

Sneh Anand

2