sample_mdape - bruno-beloff/scs_analysis GitHub Wiki
docs > software repositories > scs_analysis > commands > filtering and aggregating data
DESCRIPTION
The sample_mdape utility is used to compute Median Absolute Percent Error (MdAPE). Output is in the form of either a scalar MdAPE value, or an array of data points.
If the --errors flag is set, then the utility outputs an array of JSON documents specifying the reference, prediction and error value for each input document. Otherwise, the utility reports the computed MdAPE value.
SYNOPSIS
sample_mdape.py [-p PRECISION] [-e] [-v] REFERENCE_PATH PREDICTION_PATH
| Options | |
|---|---|
| --version | show program's version number and exit |
| -h, --help | show this help message and exit |
| -p PRECISION, --prec=PRECISION | precision (default 3 decimal places) |
| -e, --errors | output the error values |
| -v, --verbose | report narrative to stderr |
EXAMPLES - MdAPE
csv_reader.py -v collocation-data.csv | sample_mdape.py -v src.ref.PM1 exg.PM1.vB.opc-mash.Y22
EXAMPLES - ERRORS
csv_reader.py -v collocation-data.csv | sample_mdape.py -e -v src.ref.PM1 exg.PM1.vB.opc-mash.Y22 | histo_chart.py -vb -x 0 100 -c 100 ape
DOCUMENT EXAMPLE - ERRORS
[{"src.ref.PM1": 2.47, "exg.PM1.vB.opc-mash.Y22": 2.7, "ape": 9.312}, ...]