Term Deposit Prediction

Phone calls prediction to potential long-term deposit clients

1. Introduction

A term deposit is a key product offered by banks and an effective marketing campaign can play a crucial role in its sales success. This project aims to predict the outcome of phone calls made by a bank to potential long-term deposit clients. The results of this project will assist managers in prioritizing and selecting customers to contact during future marketing campaigns.

2. Problem and Dataset

The problem is approached as a binary classification problem, with the goal of predicting which clients are more likely to subscribe for a term deposit.

The dataset used in this project was collected from the UCI Machine Learning Repository. The dataset contains information on direct marketing campaigns made by a Portuguese banking institution between May 2008 and November 2010, with a total of 41,188 records and 21 fields. The classification goal is to predict whether the client will subscribe (1: yes; 0: no) to a term deposit.

3. Usage

  • Download or clone the project
  • Run the Jupyter notebook, which walks through:
    • Exploratory Data Analysis (EDA)
    • Pre-processing: SMOTE, one-hot encoding, label encoder,…
    • Models: Logistic regression, Random Forest, XGBoost, LightGBM
    • Evaluation: Accuracy, F1 score
    • Interpret model: SHAP
  • Use the trained models to predict for new observations