This project aims to compare different classification models (Logistic regression, Decision tree, and SVM) to predict airline customers' satisfaction.
- Python Version: 3.7
- Packages: pandas, numpy, sklearn, matplotlib, seaborn, selenium.
- ChromeDriver 95.0.4638.10 download.
- Airlines Customer satisfaction download.
- This data shows whether a customer is satisfied with the airlines or not based on different 9 attributes:
- Gender
- Customer Type
- Age
- Type of Travel
- Class
- Flight Distance
- Seat comfort
- Departure/Arrival time convenient
- Food and drinks
- Removed missing values.
- Sampled data for analysis and visualization.
- Visualised correlation between attributes:
- SVM is a slow algorithm, thus I used a sample of 5000 rows from the original data as opposed to 100.000 samples used for the logistic regression and Decison tree model.
- Mean Absolute Error: 0.161
- Accuracy: 0.84
- Mean Absolute Error: 0.062
- Accuracy: 0.94
- Mean Absolute Error: 0.092
- Accuracy: 0.91