classification-lib 1.0.0

Module containing various utility functions for classification tasks

classification_lib

A custom module that grants access to libraries with utility functions for performing classification tasks such as:

• Plotting precision-recall curves for multiple classifiers for easier comparisons.

• Plotting ROC curves for multiple classifiers for easier comparisons.

Installation

pip install classification_lib

Get Started

How to plot different pr-curve(s) for classifier(s) using the plot_pr_curve of this library:

from classification_lib import Plots

import numpy as np



# Instantiating the precisions and recalls for two different classifiers

classifier1_precisions = np.sort(0.791 + 0.168 * np.random.randn(794))

classifier2_precisions = np.sort(0.5 + 0.142 * np.random.randn(794))

classifier1_recalls = sorted(0.70 + 0.299 * np.random.randn(794), reverse=True)

classifier2_recalls = sorted(0.3 + 0.5 * np.random.randn(794), reverse=True)

classifier1_name = "Classifier 1"

classifier2_name = "Classifier 2"

classifiers = [(classifier1_name, classifier1_precisions, classifier1_recalls),

               (classifier2_name, classifier2_precisions, classifier2_recalls)]



# Instantiate a Plot object

plot = Plots(classifiers)



# Call the pr_curve_plot method

result = plot.plot_pr_curve()

How to plot different pr-curve(s) for classifier(s) using the plot_roc_curve of this library:

import numpy as np

from classification_lib import Plots



classifier1_false_positive_rate = sorted(0.29 + 0.248 * np.random.randn(219))

classifier2_false_positive_rate = sorted(0.7 + 0.4 * np.random.randn(219))

classifier1_true_positive_rate= sorted(0.788 + 0.2 * np.random.randn(219))

classifier2_true_positive_rate = sorted(0.6 + 0.3 * np.random.randn(219))

classifier1_name = "Classifier 1"

classifier2_name = "Classifier 2"

classifiers = [(classifier1_name, classifier1_false_positive_rate, classifier1_true_positive_rate),

               (classifier2_name, classifier2_false_positive_rate, classifier2_true_positive_rate)]



# Instantiate a Plot object

plot = Plots(classifiers)



# Call the pr_curve_plot method

result = plot.plot_roc_curve()