Airlines Customer Satisfaction Prediction

Decision Trees, Random Forest, and Gradient Boosting

Classification Model

Introduction

The airline is interested in predicting whether a future customer would be satisfied with their services given previous customer feedback about their flight experience. The airline would like you to construct and evaluate a model that can accomplish this goal. Specifically, they are interested in knowing which features are most important to customer satisfaction. The data for this activity includes survey responses from 129,880 customers. It includes data points such as class, flight distance, and in-flight entertainment, among others. Tree based models were built to predict whether or not a customer will be satisfied with their flight experience.

All the files related to this project are available at Github.com/nitin6753/Customer_Satisfaction_Prediction

Summary

Objective:

Develop a machine learning model to predict customer satisfaction with airlines services.

Tasks Performed:

Conducted extensive EDA, cleaned and structured the dataset for analysis.
Tuned hyperparameters of all models through grid search and cross-validation.
Developed an optimal DT, RF, and GB model to predict the customer satisfaction using 21 features.
Evaluated the model performance using validation sets, and then selected Gradient Boosting model as the champion by observing the evaluation metrics on test set.

Key Result Areas:

Achieved an excellent model performance with F1 score of 96.05% and other metrics including an accuracy of 95.70%, recall of 95.02%, and precision of 97.10%.
Identified key features influencing customer satisfaction: Seat comfort, Inflight entertainment, and Customer loyalty.
Demonstrated the efficacy of decision trees in predicting customer satisfaction metrics accurately.

Skills Demonstrated:

Data Wrangling, Tree Based Algorithms, Hyperparameter Tuning, Statistical Analysis, Cross-validation, Feature Importance.

Feature Importance

Model Comparison: Validation and Test data

Confusion Matrix

Home

Page updated

Google Sites

Report abuse

Airlines Customer Satisfaction Prediction

Introduction

Summary

Objective:

Develop a machine learning model to predict customer satisfaction with airlines services.

Tasks Performed:

Conducted extensive EDA, cleaned and structured the dataset for analysis.

Tuned hyperparameters of all models through grid search and cross-validation.

Developed an optimal DT, RF, and GB model to predict the customer satisfaction using 21 features.

Evaluated the model performance using validation sets, and then selected Gradient Boosting model as the champion by observing the evaluation metrics on test set.

Key Result Areas:

Achieved an excellent model performance with F1 score of 96.05% and other metrics including an accuracy of 95.70%, recall of 95.02%, and precision of 97.10%.

Identified key features influencing customer satisfaction: Seat comfort, Inflight entertainment, and Customer loyalty.

Demonstrated the efficacy of decision trees in predicting customer satisfaction metrics accurately.

Recommended:

Enhancements in seat comfort, inflight entertainment, and customer loyalty programs could boost customer satisfaction.

Skills Demonstrated:

Data Wrangling, Tree Based Algorithms, Hyperparameter Tuning, Statistical Analysis, Cross-validation, Feature Importance.