TakeConfusionMatrix 0.0.5

TakeConfusionMatrix is a tool for batched metrics calculations

TakeConfunsionMatrix

TakeConfusionMatrix is a Python package for batched Machine Learning metrics calculation and is distributed under MIT License.

Goal

The main goal of this package is to enable calculation of Machine Learning metrics for hundreds of milions of results through batch computation.

Features

The current package features are:

• Confusion Matrix
  • Image output
  • Normalized output
  • Custom labelled output
• Precision Score
• Recall Score
• F1-Score
• Accuracy Score (Exact Match Ratio)

Installation

Dependencies

TakeConfusionMatrix requires:

• Python (>= 3.8)
• Pandas (>= 1.0.4)
• scikit-learn (>= 0.23.1)

User installation

Install the 64bit version of Python, for instance from https://www.python.org/. Then run:

pip install -U TakeConfusionMatrix

VirutalEnv installation

In order to avoid potential conflicts with other packages it is strongly recommended to use a virtual environment, e.g. python3 virtualenv (see python3 virtualenv documentation) or conda environments.

To do so, install the 64bit version of Python3 if you doesn't have it yet, then run:

python3 -m venv venv
source venv/bin/activate
pip install -U TakeConfusionMatrix

NOTE: Please note that the above instructions assume a Linux-based SO. If you are using another environment, see scikit-learn installation documentation.

Usage

Here the package's features are briefly presented. For more advanced examples, please refer to the methods documentation.

Matrix computation

# Import MetricsComputation class
from take_confusion_matrix import MetricsCalculation

# Initialize class
labels = [0, 1, 2]
mc = MetricsCalculator(labels)

# Compute matrix
y_true = [0, 1, 0, 1]
y_pred = [0, 0, 0, 0]
mc.compute_matrix(y_true, y_pred)

y_true = [0, 2, 0, 2]
y_pred = [0, 0, 0, 0]
mc.compute_matrix(y_true, y_pred)

# Generate matrix
confusion_matrix = mc.generate_confusion_matrix()
print(confusion_matrix)

Normalized matrix

confusion_matrix = mc.generate_confusion_matrix(normalize=True)

Custom labelled matrix

labels = ["class_0", "class_1", "class_2"]
confusion_matrix = mc.generate_confusion_matrix(labels=labels)

Label free matrix

confusion_matrix = mc.generate_confusion_matrix(with_labels=False)

Image matrix

mc.generate_confusion_matrix(as_image=True)

Metrics computation

# Import MetricsComputation class
from take_confusion_matrix import MetricsCalculation

# Initialize class
labels = [0, 1, 2]
mc = MetricsCalculator(labels)

# Compute matrix
y_true = [0, 1, 0, 1]
y_pred = [0, 0, 0, 0]
mc.compute_matrix(y_true, y_pred)

y_true = [0, 2, 0, 2]
y_pred = [0, 0, 0, 0]
mc.compute_matrix(y_true, y_pred)

# Generate metics
metrics = mc.generate_metrics()
print(metrics)

Testing

In order to test package's features, you must download the code and change you current directory (cd) to the package's one. After that, open a terminal inside package's folder and type:

pytest

All tests are stored inside tests folders, meaning that any test folder named tests contains a test set.

Maintainer

Take's D&A Team | [e-mail]

Author

Cecília Regina Oliveira de Assis | @ceciliassis