Skip to content

In this project, I leverage machine learning models including Logistic Regression, Decision Tree, Random Forest, XGBoost, CatBoost, and LightGBM to predict customer lead scoring. I apply WOE and SHAP for feature selection and use Optuna for hyperparameter turning, aiming to identify potential lead customers effectively.

Notifications You must be signed in to change notification settings

HaiHuynh206/Lead_scoring_model

Repository files navigation

Developing machine learning models for lead scoring in customer relationship management


Introduction

  • This project aims to develop a machine learning model to automate the process of scoring potential customers (lead scoring) in Customer Relationship Management (CRM). The primary goal is to provide a tool that helps businesses better understand the potential of each customer, thereby optimizing outreach strategies and enhancing sales effectiveness.

Using the Model

  • This model allows businesses to automate the scoring of lead customers based on various criteria including online behavior, and interactions with campaigns. Below is a flow illustrating how the model operates:

Results

  • The model has achieved notable results in classifying and predicting potential customers with a high likelihood of conversion. Below are some charts and tables demonstrating the effectiveness of the model:
SHAP Train Accuracy Train F1-score Train Gini Test Accuracy Test F1-score Test Gini
CatBoost 0 0.8388 0.870 0.830 0.8371 0.866 0.816
CatBoost 1 0.787 0.793
LightGBM 0 0.8347 0.867 0.818 0.8355 0.865 0.817
LightGBM 1 0.781 0.789
  • This model helps businesses identify and focus resources on potential customers with high conversion potential, thereby optimizing sales and marketing strategies.

Hyperparameters Turning
CatBoost

LightGBM

Score

About

In this project, I leverage machine learning models including Logistic Regression, Decision Tree, Random Forest, XGBoost, CatBoost, and LightGBM to predict customer lead scoring. I apply WOE and SHAP for feature selection and use Optuna for hyperparameter turning, aiming to identify potential lead customers effectively.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published