data.imputation {postGGIR}

Data imputation for the cleaned data with annotation

Description

Data imputation for the merged ENMO data with annotation. The missing values were imputated by the average ENMO over all the valid days for each subject.

Usage

data.imputation(workdir, csvInput)

Arguments

workdir

character Directory where the output needs to be stored. Note that this directory must exist.

csvInput

character File name with or without directory for sample information in CSV format. The ENMO data will be read through read.csv(csvInput,header=1) command, and the missing values were imputated by the average ENMO over all the valid days for each subject at each time point. In this package, csvInput = flag_All_studyname_ENMO.data.Xs.csv.

Value

Files were written to the specified sub-directory, named as impu.flag_All_studyname_ENMO.data.Xs.csv, which Xs is the epoch size to which acceleration was averaged (seconds) in GGIR output. This excel file includs the following columns,

filename

accelerometer file name

Date

date recored from the GGIR part2.summary file

id

IDs recored from the GGIR part2.summary file

calender_date

date in the format of yyyy-mm-dd

N.valid.hours

number of hours with valid data recored from the part2_daysummary.csv file in the GGIR output

N.hours

number of hours of measurement recored from the part2_daysummary.csv file in the GGIR output

weekday

day of the week-Day of the week

measurementday

day of measurement-Day number relative to start of the measurement

newID

new IDs defined as the user-defined function of filename2id(), e.g. substrings of the filename

Nmiss_c9_c31

number of NAs from the 9th to 31th column in the part2_daysummary.csv file in the GGIR output

missing

"M" indicates missing for an invalid day, and "C" indicates completeness for a valid day

Ndays

number of days of measurement

ith_day

rank of the measurementday, for example, the value is 1,2,3,4,-3,-2,-1 for measurementday = 1,...,7

Nmiss

number of missing (invalid) days

Nnonmiss

number of non-missing (valid) days

misspattern

indicators of missing/nonmissing for all measurement days at the subject level

RowNonWear

number of columnns in the non-wearing matrix

NonWearMin

number of minutes of non-wearing

daysleeper

If 0 then the person is a nightsleeper (sleep period did not overlap with noon) if value=1 then the person is a daysleeper (sleep period did overlap with noon).

remove16h7day

indicator of a key qulity control output. If remove16h7day=1, the day need to be removed. If remove16h7day=0, the day need to be kept.

duplicate

If duplicate="remove", the accelerometer files will not be used in the data analysis of part5.

ImpuMiss.b

number of missing values on the ENMO data before imputation

ImpuMiss.a

number of missing values on the ENMO data after imputation

KEEP

The value is "keep"/"remove", e.g. KEEP="remove" if remove16h7day=1 or duplicate="remove" or ImpuMiss.a>0