Spaces:
Runtime error
Runtime error
File size: 4,571 Bytes
74b04a7 cc1c035 3f03a93 cc1c035 74b04a7 cc1c035 3f03a93 c79780c c710aba 3f03a93 cc1c035 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 |
---
title: MAE
emoji: 🤗
colorFrom: blue
colorTo: red
sdk: gradio
sdk_version: 3.0.2
app_file: app.py
pinned: false
tags:
- evaluate
- metric
description: >-
Mean Absolute Error (MAE) is the mean of the magnitude of difference between the predicted and actual
values.
---
# Metric Card for MAE
## Metric Description
Mean Absolute Error (MAE) is the mean of the magnitude of difference between the predicted and actual numeric values:
## How to Use
At minimum, this metric requires predictions and references as inputs.
```python
>>> mae_metric = evaluate.load("mae")
>>> predictions = [2.5, 0.0, 2, 8]
>>> references = [3, -0.5, 2, 7]
>>> results = mae_metric.compute(predictions=predictions, references=references)
```
### Inputs
Mandatory inputs:
- `predictions`: numeric array-like of shape (`n_samples,`) or (`n_samples`, `n_outputs`), representing the estimated target values.
- `references`: numeric array-like of shape (`n_samples,`) or (`n_samples`, `n_outputs`), representing the ground truth (correct) target values.
Optional arguments:
- `sample_weight`: numeric array-like of shape (`n_samples,`) representing sample weights. The default is `None`.
- `multioutput`: `raw_values`, `uniform_average` or numeric array-like of shape (`n_outputs,`), which defines the aggregation of multiple output values. The default value is `uniform_average`.
- `raw_values` returns a full set of errors in case of multioutput input.
- `uniform_average` means that the errors of all outputs are averaged with uniform weight.
- the array-like value defines weights used to average errors.
### Output Values
This metric outputs a dictionary, containing the mean absolute error score, which is of type:
- `float`: if multioutput is `uniform_average` or an ndarray of weights, then the weighted average of all output errors is returned.
- numeric array-like of shape (`n_outputs,`): if multioutput is `raw_values`, then the score is returned for each output separately.
Each MAE `float` value ranges from `0.0` to `1.0`, with the best value being 0.0.
Output Example(s):
```python
{'mae': 0.5}
```
If `multioutput="raw_values"`:
```python
{'mae': array([0.5, 1. ])}
```
#### Values from Popular Papers
### Examples
Example with the `uniform_average` config:
```python
>>> mae_metric = evaluate.load("mae")
>>> predictions = [2.5, 0.0, 2, 8]
>>> references = [3, -0.5, 2, 7]
>>> results = mae_metric.compute(predictions=predictions, references=references)
>>> print(results)
{'mae': 0.5}
```
Example with multi-dimensional lists, and the `raw_values` config:
```python
>>> mae_metric = evaluate.load("mae", "multilist")
>>> predictions = [[0.5, 1], [-1, 1], [7, -6]]
>>> references = [[0, 2], [-1, 2], [8, -5]]
>>> results = mae_metric.compute(predictions=predictions, references=references)
>>> print(results)
{'mae': 0.75}
>>> results = mae_metric.compute(predictions=predictions, references=references, multioutput='raw_values')
>>> print(results)
{'mae': array([0.5, 1. ])}
```
## Limitations and Bias
One limitation of MAE is that the relative size of the error is not always obvious, meaning that it can be difficult to tell a big error from a smaller one -- metrics such as Mean Absolute Percentage Error (MAPE) have been proposed to calculate MAE in percentage terms.
Also, since it calculates the mean, MAE may underestimate the impact of big, but infrequent, errors -- metrics such as the Root Mean Square Error (RMSE) compensate for this by taking the root of error values.
## Citation(s)
```bibtex
@article{scikit-learn,
title={Scikit-learn: Machine Learning in {P}ython},
author={Pedregosa, F. and Varoquaux, G. and Gramfort, A. and Michel, V.
and Thirion, B. and Grisel, O. and Blondel, M. and Prettenhofer, P.
and Weiss, R. and Dubourg, V. and Vanderplas, J. and Passos, A. and
Cournapeau, D. and Brucher, M. and Perrot, M. and Duchesnay, E.},
journal={Journal of Machine Learning Research},
volume={12},
pages={2825--2830},
year={2011}
}
```
```bibtex
@article{willmott2005advantages,
title={Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance},
author={Willmott, Cort J and Matsuura, Kenji},
journal={Climate research},
volume={30},
number={1},
pages={79--82},
year={2005}
}
```
## Further References
- [Mean Absolute Error - Wikipedia](https://en.wikipedia.org/wiki/Mean_absolute_error)
|