I want to make a model based on accelerometer data to recognise different activities like running, walking etc. I have a small dataset collected from my target sensor. I found another dataset with similar activities but built from a different sensor, with different sampling rate and different dynamic range. For example my sensor dynamic range is +-2g, sampling rate 20Hz other sensor dynamic range +-16g, sampling rate 100Hz

What is the best way for me to use the second dataset for training my model? I tried to downsampled the dataset B to 20Hz, and use dataset A as validation and test set, but didn't get good results. My model is based on CNN and feature extraction from 2sec window with 1sec overlap, features like mean, standard_deviation, variance, skewness, kurtosis, rms, rms_velocity, sum_of_changes and etc



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.