Difference between revisions of "Python-for-Machine-Learning/C2/Logistic-Regression-Binary-Classification/English"

Welcome

Learning Objectives

• Logistic Regression
• Binary Classification
• Multiclass Classification

System Requirements

• Ubuntu Linux OS version 24.04
• Jupyter Notebook IDE

Prerequisites

• The learner must have basic knowledge of Python.
• For pre-requisite Python tutorials, please visit this website.

Code files

• The files used in this tutorial are provided in the Code files link.
• Please download and extract the files.
• Make a copy and then use them while practicing.

Logistic Regression

• Logistic regression is a machine learning algorithm used for classification tasks.
• The goal is to predict a binary outcome (like yes/no, true/false) based on input features.

Logistic Regression

• Logistic regression uses the logit function.
• This function maps the linear combination of features to probabilities zero and one.

• Binary Classification
• Multiclass Classification

Binary classification

• Binary classification is used for modeling a binary target variable.
• The target variable has only two possible outcomes.

Multiclass classification

• Multiclass classification is an extension of binary classification.
• The target variable can have two or more possible outcomes.

Here, we analyze customers to predict if they will make a purchase in the store.

Type conda activate ml Press Enter

Activate the machine learning environment as shown.

Type jupyter notebook

Press Enter

Please navigate to the respective folder of your code file location.

Then type, jupyter space notebook and press Enter.

Double Click on LR_Binary.ipynb file

Click the LR underscore Binary dot ipynb file to open it.

Note that each cell will have the output displayed in this file.

import matplotlib.pyplot as plt

import seaborn as sns

Make sure to Press Shift and Enter to execute the code in each cell.

ads.head()

print(f"Shape of the dataset{ads.shape}")

print(ads.info())

Then we summarize the dataset, including rows, columns, and missing values using ads dot info.

sns.countplot(x='Purchased', data=ads, palette='viridis')

This attribute represents the target variable.

In the output cell, ignore the warning if you get any.

ads.head()

Let us display the first few rows of the updated data and verify.

The fit method in sklearn can't train models with string data.

Now, female is encoded as 0 and male as 1.

Have a look at the encoded data.

Highlight x = ads.drop(columns=["Purchased"])

x.head()

First we remove the target column Purchased from the ads dataset.

Then, we copy the remaining features into the variable x.

Notice that we have removed the Purchased column.

y = ads.Purchased

y.head()

The variable y has only the class labels as shown.

Highlight mms = MinMaxScaler()

First we create the instance of MinMaxScaler using the MinMaxScaler method.

The test underscore size equals 0.3 means 30% of the data is allocated for testing.

The remaining 70 percent is used for training.

Training data is used to train the model.

x underscore test and y underscore test are test features and labels.

Test data is used to evaluate the model performance.

Highlight lr = LogisticRegression()

We create an instance of LogisticRegression from the sklearn library.

It predicts the target variable for the training data.

It will return the predicted probabilities for each class.

ROC is Receiver Operating Characteristic, and AUC is Area Under the Curve.

It measures how well the model distinguishes between the two classes.

A higher score indicates better performance.

It measures how close the predicted probabilities are to the actual class labels.

A lower value indicates better model performance.

Cross Entropy Loss: 0.406

This shows the model effectively distinguishes between the classes.

The cross-entropy loss is 0.406.

It shows how well the model’s predictions match the actual labels.

For prediction we use the class of test underscore data.

We predict the classes based on test underscore data.

In the output cell, ignore the warning if you get any.

y_test_pred_proba = lr.predict_proba(x_test)[:, 1]

Cross Entropy Loss: 0.355

The model demonstrates excellent performance in distinguishing between classes.

A Cross Entropy Loss of 0.355 shows fairly accurate predictions.

conf_matrix = confusion_matrix(y_test, y_pred)

It shows how well the model is correctly classifying the instances.

However, three non-buyers are incorrectly classified as buyers.

Similarly, 24 actual buyers are correctly identified.

But 17 were misclassified as non-buyers.

A higher number in the diagonal indicated better accuracy.

This means the model performs well but struggles with some misclassifications.

We have successfully predicted if a given user will make a purchase in the store.

Summary

Only Narration

• Logistic Regression
• Binary Classification
• Multiclass Classification

In the next tutorial, we’ll learn how to implement Multiclass classification for Logistic Regression.

Assignment

As an assignment, please do the following:

Replace the  y underscore pred  code with the code as shown here.

Observe the changes in Training and Testing accuracy.

Assignment Solution

binary.png

FOSSEE Forum

Machine Learning, visit the FOSSEE forum and post your question.

Thank You

Thanks for joining.