Regression Metrics Calculation Made easy.
Project description
Regression Metrics
Installation
To install the package from the PyPi repository you can execute the following command:
pip install regressionmetrics
If you prefer, you can clone it and run the setup.py file. Use the following commands to get a copy from GitHub and install all dependencies:
git clone https://github.com/ashishpatel26/regressionmetrics.git
cd regressionmetrics
pip install .
Metrics | Full Form | Interpretation | Sklearn | Keras |
---|---|---|---|---|
mae | Mean Absolute Error | Smaller is better (Best value is 0) | ☑️ | ☑️ |
mse | Mean Sqaured Error | Smaller is better(Best value is 0) | ☑️ | ☑️ |
rmse | Root Mean Square Error | Smaller is better(Best value is 0) | ☑️ | ☑️ |
rmsle | Root Mean Square Log Error | Smaller is better(Best value is 0) | ☑️ | ☑️ |
rmsle_with_negval | Root Mean Square Log Error with neg. value | Smaller is better(Best value is 0) | ☑️ | |
r2 score | coefficient of determination | Best possible score is 1 | ☑️ | ☑️ |
Adjusted r2 score | Adjusted R2 score | Best possible score is 1 | ☑️ | ☑️ |
mape | Mean Absolute Percentage Error | Smaller is better(Best value is 0) | ☑️ | ☑️ |
msle | Mean Sqaured Logarithm Error | Smaller is better(Best value is 0) | ☑️ | ☑️ |
smape | Symmetric mean absolute percentage error | Smaller is better(Best value is 0) | ☑️ | |
nrmse | Normalized Root Mean Square Error. | ☑️ | ☑️ | |
nrmsle | Normalized Root Mean Squared Logarithmic Error | ☑️ | ||
medianAE | Median Absolute Error | Smaller is better(Best value is 0) | ☑️ | |
mre | Mean Relative Error | Smaller is better(Best value is 0) | ☑️ | |
maape | Mean Arctangent Absolute Percentage Error | Smaller is better(Best value is 0) | ☑️ | |
nse | Nash-Sutcliffe Efficiency Coefficient | Larger is better (Best = 1) | ☑️ | |
willmott_index_of_agreement | Willmott Index | Larger is better (Best = 1) | ☑️ |
Usage
Usage with scikit learn :
from regressionmetrics.metrics import *
y_true = np.array([3, 0.5, 2, 7])
y_pred = np.array([2.5, 0.0, 2, -8])
print("R2Score: ",r2(y_true, y_pred))
print("Adjusted_R2_Score:",adj_r2(y_true, y_pred))
print("RMSE:", rmse(y_true, y_pred))
print("MAE:",mae(y_true, y_pred))
print("RMSLE with Neg Value:", rmsle_with_negval(y_true, y_pred))
print("MSE:", mse(y_true, y_pred))
print("MAPE: ", mape(y_true, y_pred))
Usage with Tensorflow keras:
from regressionmetrics.keras import *
import pandas as pd
import numpy as np
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.boston_housing.load_data(path="boston_housing.npz", test_split=0.2, seed=113)
model = keras.Sequential([
layers.Dense(64, activation='relu', input_shape=(x_train.shape[1],)),
layers.Dense(64, activation='relu'),
layers.Dense(1)
])
model.compile(optimizer='rmsprop', loss='mse', metrics=[r2, mae, mse, rmse, mape, rmsle, nrmse])
model.fit(x_train, y_train, epochs=10, batch_size=32, validation_data=(x_test, y_test))
Epoch 1/10
1/13 [=>............................] - ETA: 7s - loss: 1574.7567 - r2: 0.6597 - mae: 37.1803 - mse: 1574.7567 - rmse: 37.1802 - mape: 159.261313/13 [==============================] - 1s 15ms/step - loss: 270.0653 - r2: 0.9472 - mae: 11.5427 - mse: 270.0653 - rmse: 11.5427 - mape: 57.3519 - rmsle: 0.6445 - nrmse: 0.5735 - val_loss: 88.6351 - val_r2: 0.9727 - val_mae: 6.6028 - val_mse: 88.6351 - val_rmse: 6.6028 - val_mape: 29.6502 - val_rmsle: 0.3161 - val_nrmse: 0.2965
Epoch 2/10
1/13 [=>............................] - ETA: 0s - loss: 74.6623 - r2: 0.9913 - mae: 5.5958 - mse: 74.6623 - rmse: 5.5958 - mape: 25.3655 - rmsl13/13 [==============================] - 0s 3ms/step - loss: 87.1876 - r2: 0.9856 - mae: 6.9466 - mse: 87.1876 - rmse: 6.9466 - mape: 33.4256 - rmsle: 0.3057 - nrmse: 0.3343 - val_loss: 81.7884 - val_r2: 0.9712 - val_mae: 6.6424 - val_mse: 81.7884 - val_rmse: 6.6424 - val_mape: 28.8687 - val_rmsle: 0.3334 - val_nrmse: 0.2887
Epoch 3/10
1/13 [=>............................] - ETA: 0s - loss: 41.2790 - r2: 0.9722 - mae: 5.3798 - mse: 41.2790 - rmse: 5.3798 - mape: 28.7497 - rmsl13/13 [==============================] - 0s 3ms/step - loss: 103.6462 - r2: 0.9825 - mae: 7.1041 - mse: 103.6462 - rmse: 7.1041 - mape: 34.6278 - rmsle: 0.3231 - nrmse: 0.3463 - val_loss: 71.7539 - val_r2: 0.9769 - val_mae: 6.1455 - val_mse: 71.7539 - val_rmse: 6.1455 - val_mape: 27.5078 - val_rmsle: 0.2893 - val_nrmse: 0.2751
Epoch 4/10
1/13 [=>............................] - ETA: 0s - loss: 113.6758 - r2: 0.9917 - mae: 6.6575 - mse: 113.6758 - rmse: 6.6575 - mape: 20.8683 - rm13/13 [==============================] - 0s 3ms/step - loss: 88.1601 - r2: 0.9823 - mae: 6.8479 - mse: 88.1601 - rmse: 6.8479 - mape: 32.5867 - rmsle: 0.3080 - nrmse: 0.3259 - val_loss: 63.3707 - val_r2: 0.9829 - val_mae: 6.0845 - val_mse: 63.3707 - val_rmse: 6.0845 - val_mape: 33.1628 - val_rmsle: 0.2747 - val_nrmse: 0.3316
Epoch 5/10
1/13 [=>............................] - ETA: 0s - loss: 85.8188 - r2: 0.9893 - mae: 7.0097 - mse: 85.8188 - rmse: 7.0097 - mape: 34.8362 - rmsl13/13 [==============================] - 0s 3ms/step - loss: 82.3233 - r2: 0.9860 - mae: 6.5795 - mse: 82.3233 - rmse: 6.5795 - mape: 32.5198 - rmsle: 0.3105 - nrmse: 0.3252 - val_loss: 74.4783 - val_r2: 0.9813 - val_mae: 6.8936 - val_mse: 74.4783 - val_rmse: 6.8936 - val_mape: 41.9492 - val_rmsle: 0.3067 - val_nrmse: 0.4195
Epoch 7/10
1/13 [=>............................] - ETA: 0s - loss: 105.6430 - r2: 0.9658 - mae: 9.4737 - mse: 105.6430 - rmse: 9.4737 - mape: 53.0854 - rm13/13 [==============================] - 0s 3ms/step - loss: 76.0740 - r2: 0.9856 - mae: 6.4234 - mse: 76.0740 - rmse: 6.4234 - mape: 31.8728 - rmsle: 0.2828 - nrmse: 0.3187 - val_loss: 104.1779 - val_r2: 0.9679 - val_mae: 7.5539 - val_mse: 104.1779 - val_rmse: 7.5539 - val_mape: 30.9401 - val_rmsle: 0.3692 - val_nrmse: 0.3094
Epoch 8/10
1/13 [=>............................] - ETA: 0s - loss: 100.0114 - r2: 0.9833 - mae: 6.8492 - mse: 100.0114 - rmse: 6.8492 - mape: 27.9621 - rm13/13 [==============================] - 0s 4ms/step - loss: 68.4268 - r2: 0.9892 - mae: 5.9540 - mse: 68.4268 - rmse: 5.9540 - mape: 29.7586 - rmsle: 0.2623 - nrmse: 0.2976 - val_loss: 171.7968 - val_r2: 0.9412 - val_mae: 10.5855 - val_mse: 171.7968 - val_rmse: 10.5855 - val_mape: 47.9010 - val_rmsle: 0.7561 - val_nrmse: 0.4790
Epoch 9/10
1/13 [=>............................] - ETA: 0s - loss: 291.8670 - r2: 0.9725 - mae: 13.9899 - mse: 291.8670 - rmse: 13.9899 - mape: 61.3658 - 13/13 [==============================] - 0s 3ms/step - loss: 92.3889 - r2: 0.9796 - mae: 6.8932 - mse: 92.3889 - rmse: 6.8932 - mape: 33.2856 - rmsle: 0.3333 - nrmse: 0.3329 - val_loss: 67.2208 - val_r2: 0.9808 - val_mae: 5.8498 - val_mse: 67.2208 - val_rmse: 5.8498 - val_mape: 26.4504 - val_rmsle: 0.2680 - val_nrmse: 0.2645
Epoch 10/10
1/13 [=>............................] - ETA: 0s - loss: 97.0853 - r2: 0.9923 - mae: 5.9866 - mse: 97.0853 - rmse: 5.9866 - mape: 24.9878 - rmsl13/13 [==============================] - 0s 3ms/step - loss: 78.3823 - r2: 0.9856 - mae: 6.5958 - mse: 78.3823 - rmse: 6.5958 - mape: 32.8136 - rmsle: 0.3025 - nrmse: 0.3281 - val_loss: 69.5314 - val_r2: 0.9787 - val_mae: 6.8302 - val_mse: 69.5314 - val_rmse: 6.8302 - val_mape: 37.3933 - val_rmsle: 0.2974 - val_nrmse: 0.3739
:smiley: Thanks for reading and forking.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
regressionmetrics-1.4.0.tar.gz
(10.8 kB
view details)
Built Distribution
File details
Details for the file regressionmetrics-1.4.0.tar.gz
.
File metadata
- Download URL: regressionmetrics-1.4.0.tar.gz
- Upload date:
- Size: 10.8 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.4.2 importlib_metadata/4.8.1 pkginfo/1.7.1 requests/2.26.0 requests-toolbelt/0.9.1 tqdm/4.62.1 CPython/3.8.5
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | e3b0e637565b2974bd8933b9932c26273e8b4e0de771fbc4f54fc1d97e6588cb |
|
MD5 | fd6e8fa8fee13a562a9733b1b78a6775 |
|
BLAKE2b-256 | 681c62023b22db7ce095febff4dd179e888dc71996db2699b29be60efabec66c |
File details
Details for the file regressionmetrics-1.4.0-py3-none-any.whl
.
File metadata
- Download URL: regressionmetrics-1.4.0-py3-none-any.whl
- Upload date:
- Size: 9.0 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.4.2 importlib_metadata/4.8.1 pkginfo/1.7.1 requests/2.26.0 requests-toolbelt/0.9.1 tqdm/4.62.1 CPython/3.8.5
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 77e9265f68730ae6e6b8f9df67db09fb48ef6358d09a24deea6c2a2eff946441 |
|
MD5 | ee21f63427619be4638af7d907b6232a |
|
BLAKE2b-256 | 095eb69159d5bfbe3255ef5265285edf811b724dec8a4fcbada7461ee331e640 |