Bank Marketing Machine Learning Analysis

This repository contains a machine learning project analyzing the UCI Bank Marketing dataset. The objective is to compare the performance of traditional machine learning models (Logistic Regression, Random Forest) with a feed-forward neural network implemented in PyTorch. The task is to predict whether a customer will subscribe to a term deposit based on demographic, financial, and campaign-related features.

Project Overview

Financial institutions rely on targeted marketing to identify potential customers. This project evaluates how classical machine learning methods compare to a neural network in predicting customer subscription outcomes. The work includes exploratory data analysis, preprocessing, baseline model development, neural network training, and a final comparison of results.

Repository Structure

bank-marketing-ml/
│
├── README.md
├── report/
│   └── project_report.pdf
│
├── data/
│   ├── raw/
│   │   ├── bank-full.csv
│   │   ├── bank.csv
│   │   └── bank-names.txt
│   └── processed/
│       └── bank_processed_unbalanced.csv
│
├── notebooks/
│   ├── 01_exploration.ipynb
│   ├── 02_preprocessing.ipynb
│   ├── 03_baseline_models.ipynb
│   └── 04_neural_network.ipynb
│
├── src/
│   ├── preprocess.py
│   ├── train_baseline.py
│   ├── train_nn.py
│   └── utils.py
│
├── models/
│   ├── logistic_regression.pkl
│   ├── random_forest.pkl
│   └── neural_net.pt
│
├── results/
│   ├── metrics/
│   └── figures/
│
└── requirements.txt

Methods

Baseline Models:

• Logistic Regression (scikit-learn)
• Random Forest (scikit-learn)

Neural Network:

• Fully connected feed-forward network implemented in PyTorch
• Standardization and feature encoding applied prior to training
• Trained using Adam optimizer and binary cross-entropy loss

Evaluation

Models are compared using:

• Accuracy
• Precision, recall, and F1-score
• Confusion matrices
• Training and validation loss (for the neural network)

The final report, located in the report/ directory, provides full details, analysis, and discussion of results.

How to Run

Install requirements:

pip install -r requirements.txt

Run preprocessing:

python src/preprocess.py

Train baseline models:

python src/train_baseline.py

Train the neural network:

python src/train_nn.py

Data Source

Moro, Sérgio, Paulo Cortez, and Paulo Rita. “Bank Marketing Data Set.” UCI Machine Learning Repository, 2014.

Source Code

All code used for data processing, model training, and evaluation is included in this repository.