Skip to content

CodeBook.md

Latest commit

 

History

History
50 lines (36 loc) · 2.08 KB

CodeBook.md

File metadata and controls

50 lines (36 loc) · 2.08 KB

#CodeBook

##Introduction This file describes the variables in new_data.txt which is produced at step 5 in the course project. This file also describes briefly the process to produce that file.

##Raw Data Raw data come from this (dataset)[http://archive.ics.uci.edu/ml/datasets/Human+Activity+Recognition+Using+Smartphones]. This dataset is data collect from accelerometer and gyroscope in Samsung Galaxy S II

##Data new_data.txt is a 180x81 matrix. Each row consists values of 79 variables (more info below) for each subject and activity. Value in each 79 variable (except subject and activity) of each row is the mean. subject is integer which represents the person performing the acitivity. activiy consists 6 types of activity the person can perform: walking, walking upstairs, walking downstairs, stiing, standing, laying.

##Variables There are 81 variables in new_data.txt. 79 variables are the combination of those signal:

  • timeBodyAcc-XYZ
  • timeGravityAcc-XYZ
  • timeBodyAccJerk-XYZ
  • timeBodyGyro-XYZ
  • timeBodyGyroJerk-XYZ
  • timeBodyAccMag
  • timeGravityAccMag
  • timeBodyAccJerkMag
  • timeBodyGyroMag
  • timeBodyGyroJerkMag
  • freqBodyAcc-XYZ
  • freqBodyAccJerk-XYZ
  • freqBodyGyro-XYZ
  • freqBodyAccMag
  • freqBodyAccJerkMag
  • freqBodyGyroMag
  • freqBodyGyroJerkMag

with three measurements

  • mean: mean value
  • std: standard deviation
  • meanFreq: Weighted average of the frequency components to obtain a mean frequency

if the signal has "-XYZ" suffix that means there is separate variable for each axis (X, Y, Z)

NOTE:

  • Combination does not mean there are exact pairs of signal-measurement.
  • meanFreq is included for any further usage may involve

##Process Raw data contains training and testing datasets. Those datasets are combined to produce the dataset of 10299x561. After some transformation (adding subject and activity column) and removing duplicate columns, we have 10299x479 dataset. From that a subset 10299x81 dataset was extracted, with 81 variables describe above. And then using average in each varible for each activity and each subject we got the 180x81 dataset above.