# ---------------------------------------------------------------------
#
# R functions to manipulate gait data from TUG tests
#
# Author: Moacir A. Ponti
# -> in collaboration with: Patricia Bet; Paula Castro
#
# There are two types of files in this repository:
#
# 1. Code
#
# - gaitAnalysis.R : contains several functions to 
#                  read text files with accelerometer data
#		   and performing feature extraction
#
# - featureROCAnalysis.R: contains functions to allow
#		   performing ROC curve analysis using the
#		   extracted features or any other variable
#
#   in order to load the code files use following R commands:
#   > source('gaitAnalysis.R')
#   > source('featureROCAnalysis.R')
#
# - plotSS_fusionROC.R: code used to plot ROC curves and
#		   Sensitivity vs Specificity analysis that
#		   are included in the paper
#
# 2. Object files:
#
# - featCompleteTUG_44.rds: R object containing all features, 
#		   anonymized IDs and the group labels for each
#		   participant
#
# - featSignificantTUG_7.rds: R object containing only the 
#		   features that were found to be significant with
#		   respect to the different groups, containing also
#                  the group labels for each participant
#
# - featSecondsTUG_3.rds: R object containing the time spent by
#		   the participants while performing each TUG
#
# - featFusionFeatsDists_2.rds: result of the fusion of the Frequency 
#		   and Distance-based features.
#
# - featFusionTUGS_1.rds: result of the fusion of the TUG seconds#
#
#   in order to load some feature data, use the following R command:
#   > feats <- readRDS("file.rds")
#
#
# This repository was made public in order to allow for
#	reproducibility of the results reported in the paper
#
#	M. Ponti, P. Bet, C.L. Oliveira and P.C. Castro
#	"Better than Counting Seconds: Identifying Fallers among 
#	 Healthy Elderly using Fusion of Accelerometer Features and 
#	 Dual-Task Timed Up and Go"
#
#	the paper is currently accepted and in production at PLoS ONE
#
#	a preprint version can be found in the following arXiV link:
#       OBS: note that the arXiV paper can different from the one
#	     submitted
#	https://arxiv.org/abs/1609.05339
#
# ------------------------------------------------------------------