This repository contains all the code related to my Final Year Project which is about recognizing human activities using Smartphone and Smartwatch.
An important contribution is the exploration on how using mean deducted sensory data can significantly improve the testing results, especially the Leave-One-Person-Out (LOPO) testing. To my own knowledge, there has not been any other work using a similar technique to a Human Activity Recognition problem.
README
files are provided only for the client
and server
folders as only these two involve non-trivial setups. There is no guide for the Smartphone and Smartwatch as it can be found easily from the internet.
I suggest anyone who are interested in using the code to read the code one by one and understand the logic flow. I have given my best effort to write the README
files, but writing very detailed step-by-step documents are too time consuming and I do not think it is necessary. Reading the code should be the best way to understand everything.
Please contact me through this GitHub account if you would like to get the full raw data I have collected. I removed it from this repository because it is too big to download and it is harder to maintain the GitHub repo due to large file limitation imposed by GitHub.