NAME
heri-stat - calculates precision, recall, F1 and some other things
SYNOPSIS
heri-stat [-R] [-mrca] [-u label] [-t threshold]
outcomes_file predictions_file
heri-stat -1 [-R] [-mrca] [-u label] [-t threshold]
[files...]
heri-stat -g mode [-R] [-xy] outcomes_file predictions_file
heri-stat -1 -g mode [-R] [files...]
heri-stat [-h]
DESCRIPTION
The first and second types of heri-stat invocation takes classification dataset and predictions on input, and calculate precision, recall and F1. Unless option -1 was applied, heri-stat reads correct classes from outcomes_file (one class per line) and predicted classes from predictions_file (one class per line). It is allowed for predictions_file to contain two tokens per line. The first one is a predicted class, the second one is a score, e.g. probability.
The third and forth type of heri-stat invocation takes regression outcomes and predictions on input (one value per line) and calculate mean absolute error (MAE), mean squared error (MSE) and/or mean absolute error (MAE).
If -1 was applied, two or three tokens per line are expected on input: correct value (or class for classification), predicted value (or class), and optional score.
OPTIONS
- -h, --help
-
Display help information.
- -R, --raw
-
Raw tab-separated output.
- -m, --micro-avg
-
Disable micro averaged P/R/F1 output.
- -r, --macro-avg
-
Disable macro averaged P/R/F1 output.
- -c, --per-class
-
Disable output of per-class statistics.
- -a, --accuracy
-
Disable output of accuracy.
- -1, --single
-
2 or 3 tokens per line are expected on input.
- -u, --unclassified label
-
Set the label for "unclassified" object. If specified, micro-averaged P/R/F1 is calculated instead of accuracy. Also, statistics for this class is not calculated.
- -t, --treshold value
-
Consider classes with score less than the specified value as unclassified.
- -g, --regression mode
-
Regression outcomes and predictions are expected on input. If mode contains symbol a, MAE is output, s -- MSE, and r -- RMSE is output.
HOME
<http://github.com/cheusov/herisvm>