There are a lot of factors that are considered for deciding which parent gets the custody of their child post a divorce. Many of these factors that are considered are sensitive. So, while building a prediction model for the same, it is important that the model doesn’t defy privacy. So, a differentially private classification model for predicting the custody of a child is implemented.
Dataset Used: Divorced_Data.csv -> Dataset used for training
Original Dataset: https://www.kaggle.com/aagghh/divorcemarriage-dataset-with-birth-dates
Code: EAD_Assignment.ipynb -> Notebook containing the Training Code
Training Pickle File: SVMFinal.pkl -> Pickle file of trained model
Report: EAD_Report.pdf -> Assignment Report