/samsung-prediction

Predicting user activity based on smartphone data

Primary LanguageR

Samsung activity data analysis

My data analysis of the Samsung dataset for the Data Analysis course taught by Jeff Leek on Coursera.

Here I attempt to use combine SVD for dimensionality reduction and SVM for classification in order to predict a user's activity (e.g. standing, sitting, walking) based on smartphone accelerometer and gyroscope data.

View my report here.

To reproduce the exact analyses presented in the report, run source('00-download.r') to get the dataset then source('main.r') from the R command line.

Required packages

  • reshape
  • ggplot2

License

This work is licensed under the Creative Commons Attribution-ShareAlike 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-sa/3.0/ or send a letter to Creative Commons, 444 Castro Street, Suite 900, Mountain View, California, 94041, USA.