Week 3/readme.md

Machine Learning Zoomcamp 2025 - Homework 3

---

Homework 3: Machine Learning for Classification

This repository contains solutions for Homework 3 of Machine Learning Zoomcamp 2025, focused on classification tasks using the Bank Marketing dataset.

---

📂 Project Overview

• Dataset: Bank Marketing Dataset
• Target variable: converted (whether the client signed up)
• Objective: Data preprocessing, exploratory analysis, feature selection, and training logistic regression models (regularized and unregularized).

Tech Stack:

• Python 3.11 – core programming language
• Pandas – data manipulation
• NumPy – numerical operations
• Scikit-Learn – machine learning models, feature selection, evaluation
• Jupyter Notebook – interactive coding and documentation

---

🔹 Questions & Answers

Question	Task	Answer
1	Mode of industry	retail
2	Biggest correlation (numerical features)	annual_income and interaction_count
3	Biggest mutual information (categorical features)	lead_source
4	Logistic regression validation accuracy	0.74
5	Least useful feature (feature elimination)	lead_score
6	Best C value for regularized logistic regression	1

---

📌 Approach / Key Steps

1. Data Cleaning & Preparation

   • Filled missing values: categorical → 'NA', numerical → 0.0
   • Verified feature types and correlations

2. Exploratory Analysis

   • Mode of categorical variables
   • Correlation matrix for numerical features

3. Feature Selection

   • Calculated mutual information for categorical variables using mutual_info_score
   • Identified least useful features via feature elimination

4. Model Training

   • Logistic Regression with one-hot encoded categorical variables
   • Regularized logistic regression with hyperparameter tuning (C values)

---

📈 Results

• Baseline logistic regression accuracy: 0.74
• Least useful feature: lead_score
• Best regularization parameter C: 1

---

⚙ How to Run

1. Clone the repository:

   git clone https://github.com/yourusername/ml-zoomcamp-hw3.git
   

2. Install requirements:

   pip install -r requirements.txt
   

3. Open the Jupyter Notebook and run cells sequentially:

   jupyter notebook
   

---

📚 References

• Bank Marketing Dataset
• Scikit-Learn Documentation
• Pandas Documentation
• NumPy Documentation
• Jupyter Notebook Documentation

---