Neck_motion_classifier

This repository contains the main code developed in Google Colaboratory for a neck motion classifier towards a future implementation in real-time. The code is organized as mentioned in the paper submitted to Optical Fiber Technology and the dataset was generated in four sessions. The code was storaged in a zip file due to its weight. Moreover, the article shows the results of eight sessions. For the evaluation using var/RMS, Hjorth parameters or both, you need to uncomment the labels for dataframe creation that you wanna obtain as well as change the number of the features in the "features" function. If you need to plot something, it's possible to uncomment some parts of the code to see the graphs that plot or validate the system. In addition, the "metrics" file includes the analysis made in the article considering accuracy and AUC for each classifier and using or not the low-pass Butterworth filter.