👋 Introduction

Hey there, fellow tech enthusiasts! As a B.Tech Computer Science student, my first dive into machine learning models was a mix of excitement and a bit of intimidation. Recently, I tackled Logistic Regression, a powerful and easy-to-understand classification algorithm. Join me in this blog as I share my journey, insights, and how I wrapped my head around the math and code behind it. Let's explore this together!

---

📌 What is Logistic Regression?

Logistic regression might sound like it's about regression, but it's actually used for classification tasks. This means it's great for predicting things that have two possible outcomes. Imagine you're trying to guess if someone will buy a product or not. It's like asking, "Will they buy it? Yes or No?" Or think about sorting emails into "Spam" or "Not Spam."

The magic behind logistic regression is the sigmoid function. This function helps turn any number into a value between 0 and 1. It's like having a dial that adjusts predictions to fit between these two extremes.

For example, let's say you're trying to predict if it will rain tomorrow. The sigmoid function might give you a number like 0.8, which means there's an 80% chance of rain. If it gives you 0.2, there's only a 20% chance. This makes it easy to decide between two options, like "Rain" or "No Rain."

In the graph above, you can see how the sigmoid function curves, turning any input into a value between 0 and 1. This is what helps logistic regression make clear yes-or-no predictions.

---

📂 Dataset Used

For this project, I used a synthetic dataset of 50 samples, with the following columns:

• Age (in years)

• Salary (in $1000s)

• Purchased (0 = No, 1 = Yes)

This is the kind of data you might find in real-world marketing analysis.

---

🧪 Steps I Followed

✅ Step 1: Import Libraries

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score, confusion_matrix
import matplotlib.pyplot as plt

✅ Step 2: Load and Prepare Data

df = pd.read_csv("Customer_Purchase.csv")
X = df[['Age', 'Salary']]
y = df['Purchased']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)

✅ Step 3: Train Model

model = LogisticRegression()
model.fit(X_train, y_train)
y_pred = model.predict(X_test)

✅ Step 4: Evaluate

print("Accuracy:", accuracy_score(y_test, y_pred))
print("Confusion Matrix:\n", confusion_matrix(y_test, y_pred))

✅ Step 5: Plot Dataset (Scatter Plot)

To understand how the data is distributed:

import matplotlib.pyplot as plt

plt.figure(figsize=(8,6))
plt.scatter(df[df.Purchased == 0]['Age'], df[df.Purchased == 0]['Salary'], color='red', label='Not Purchased')
plt.scatter(df[df.Purchased == 1]['Age'], df[df.Purchased == 1]['Salary'], color='green', label='Purchased')
plt.xlabel('Age')
plt.ylabel('Salary (in $1000s)')
plt.title('Customer Purchase Data')
plt.legend()
plt.grid(True)
plt.show()

Customer Purchase Data Scatter Plot

---

📊 Insights Gained

• Logistic regression performs surprisingly well even with small datasets.

• Features like Age and Salary had clear influence on purchasing behavior.

• Visualizations helped me intuitively understand classification boundaries.

---

💡 Key Takeaways

• Logistic regression is a great starting point for beginners in ML.

• It’s interpretable and doesn’t require too much computing power.

• Feature scaling and visualization are crucial to performance and understanding.

---

👨‍🎓 Final Thoughts

Working on this project was a milestone for me as a B.Tech student. Not only did I implement my first machine learning model, but I also learned how to evaluate and explain it.

If you're new to ML, I highly recommend starting with logistic regression. It's the perfect gateway into the world of machine learning.

---

---

📄 Suggested Reading

• Logistic Regression on Scikit-Learn Docs

Thanks for reading — and happy coding! 🚀