Using Machine Learning Models to Improve the Performance of Activity Recognition for Specific Users

What is it about?

We explored certain machine learning experiments for recognising, which activity (like sitting, washing hands) a subject is performing, for instance based on smartwatch data. We analyse to what extent the performance can be increased for a specific subject when we use data from that same subject for training.

Featured Image

Why is it important?

The results show that already few data samples from a specific subject can help the recognition rates for said individual. Also, our combination of machine learning algorithms is easy-to-apply and needs limited resources compared to other approaches.

Perspectives