Accelerometer raw data of 7606 box punches - annotated on a smartwatch

---

Short Description

This dataset contains all annotated box punches. It consists of:

• 4 different classes (frontal punch, side-cut, hook, no-action)
• left / right orientated
• of 8 different subjects (boxers)

This dataset is part of the "smartPunch" project were box punch data measured from a smartwatch app is classified by machine learning algorithms like decision trees, SVMs, logistical regression and k-nearest-neighbor.

---

Gettings started

See the detailed description of this dataset and the structure at the github repository for that dataset. Its available at:
See also the other project files like the Database, server application and smartwatch app files and - the machine learning files. All the stuff is delivered with some tutorials which will help you to understand the hole project and challenge the box punch classification using the smartwatch app and the machine learning algorithms.
__Happy Coding :-)__