Inspiration
Our inspiration for this project was Western AI's DataQuest 2023.
What it does
Our project creates a machine learning model which extracts the features of importance in Brescia Norton's cancellations. Through this, our model is able to accurate ppredict future cancellations and thus allow us to create a business plan to improve customer retention.
How we built it
Technologies used:
- Python
- Jupyter
- Sklearn library
- pandas
- matplotlib from pyplot
- Excel and csv files
Design Flow:
- Built two ML models: Decision Tree and Random Forest Model
- Trained and tested both with different accuracy measurements (AUC and cfl)
- Chose the best performing model for our project : Decision Tree.
- Showcased the performance by coding functions to graph the ROC curve and the Precision vs Recall curve.
Challenges we ran into
- Cleansing data and normalization: We solved this through both one-hot encoding and mapping lambda functions to the pandas DataFrame.
Accomplishments that we're proud of
Our model performs at an AUC rate of 0.947.
What we learned
ML models are challenging but fun.
What's next for Data Quest 2023: Brescia Norton Cancellation ML Model: hopefully a winning prize!
Built With
- jupyter
- python
Log in or sign up for Devpost to join the conversation.