Bank Customer Churn Prediction Using Machine Learning
Bank Customer Churn Prediction Using Machine Learning
Bank Customer Churn Prediction Using Machine Learning
Bank Customer Churn Prediction Using Machine Learning
Machine Learning
Data Science
Python
Binary Classification with a Bank Churn Dataset
The primary objective of this project was to develop a predictive model capable of accurately determining whether a bank customer will maintain or close their account.
Kaggle Playground Series - Season 4, Episode 1
Competition Sponsor: Google LLC
Competition Website: https://www.kaggle.com/competitions/playground-series-s4e1
The leaderboard rank achieved is #15. The final entry was TOP 0.05%.
This project was specifically designed to participate in Kaggle Playground Series S4E1. In today's dynamic business environment, customer churn poses a substantial financial risk for companies. With the help of advanced data analysis, feature engineering, and machine learning techniques, this project aims to provide valuable insights and predictive capabilities to banks to mitigate customer churn.
Data:
train.csv - the training dataset; Exited is the binary target
test.csv - the test dataset; your objective is to predict the probability of Exited
sample_submission.csv - a sample submission file in the correct format
The submission file should follow the specified format for each ID in the test set. In this format, we were required to predict the probability of the target variable "Exited".
Approach:
The project began with a comprehensive data exploration phase, followed by detailed feature engineering. Several models, including ensemble models and a Neural Network, were trained and evaluated. The model with the highest AUC score was selected for further refinement and improvement. Finally, a submission file was created and submitted as the final outcome of the project.
Project Files:
Machine Learning
Data Science
Python
Binary Classification with a Bank Churn Dataset
The primary objective of this project was to develop a predictive model capable of accurately determining whether a bank customer will maintain or close their account.
Kaggle Playground Series - Season 4, Episode 1
Competition Sponsor: Google LLC
Competition Website: https://www.kaggle.com/competitions/playground-series-s4e1
The leaderboard rank achieved is #15. The final entry was TOP 0.05%.
This project was specifically designed to participate in Kaggle Playground Series S4E1. In today's dynamic business environment, customer churn poses a substantial financial risk for companies. With the help of advanced data analysis, feature engineering, and machine learning techniques, this project aims to provide valuable insights and predictive capabilities to banks to mitigate customer churn.
Data:
train.csv - the training dataset; Exited is the binary target
test.csv - the test dataset; your objective is to predict the probability of Exited
sample_submission.csv - a sample submission file in the correct format
The submission file should follow the specified format for each ID in the test set. In this format, we were required to predict the probability of the target variable "Exited".
Approach:
The project began with a comprehensive data exploration phase, followed by detailed feature engineering. Several models, including ensemble models and a Neural Network, were trained and evaluated. The model with the highest AUC score was selected for further refinement and improvement. Finally, a submission file was created and submitted as the final outcome of the project.
Project Files:
Machine Learning
Data Science
Python
Binary Classification with a Bank Churn Dataset
The primary objective of this project was to develop a predictive model capable of accurately determining whether a bank customer will maintain or close their account.
Kaggle Playground Series - Season 4, Episode 1
Competition Sponsor: Google LLC
Competition Website: https://www.kaggle.com/competitions/playground-series-s4e1
The leaderboard rank achieved is #15. The final entry was TOP 0.05%.
This project was specifically designed to participate in Kaggle Playground Series S4E1. In today's dynamic business environment, customer churn poses a substantial financial risk for companies. With the help of advanced data analysis, feature engineering, and machine learning techniques, this project aims to provide valuable insights and predictive capabilities to banks to mitigate customer churn.
Data:
train.csv - the training dataset; Exited is the binary target
test.csv - the test dataset; your objective is to predict the probability of Exited
sample_submission.csv - a sample submission file in the correct format
The submission file should follow the specified format for each ID in the test set. In this format, we were required to predict the probability of the target variable "Exited".
Approach:
The project began with a comprehensive data exploration phase, followed by detailed feature engineering. Several models, including ensemble models and a Neural Network, were trained and evaluated. The model with the highest AUC score was selected for further refinement and improvement. Finally, a submission file was created and submitted as the final outcome of the project.
Project Files:
Machine Learning
Data Science
Python
Binary Classification with a Bank Churn Dataset
The primary objective of this project was to develop a predictive model capable of accurately determining whether a bank customer will maintain or close their account.
Kaggle Playground Series - Season 4, Episode 1
Competition Sponsor: Google LLC
Competition Website: https://www.kaggle.com/competitions/playground-series-s4e1
The leaderboard rank achieved is #15. The final entry was TOP 0.05%.
This project was specifically designed to participate in Kaggle Playground Series S4E1. In today's dynamic business environment, customer churn poses a substantial financial risk for companies. With the help of advanced data analysis, feature engineering, and machine learning techniques, this project aims to provide valuable insights and predictive capabilities to banks to mitigate customer churn.
Data:
train.csv - the training dataset; Exited is the binary target
test.csv - the test dataset; your objective is to predict the probability of Exited
sample_submission.csv - a sample submission file in the correct format
The submission file should follow the specified format for each ID in the test set. In this format, we were required to predict the probability of the target variable "Exited".
Approach:
The project began with a comprehensive data exploration phase, followed by detailed feature engineering. Several models, including ensemble models and a Neural Network, were trained and evaluated. The model with the highest AUC score was selected for further refinement and improvement. Finally, a submission file was created and submitted as the final outcome of the project.