1. What is the data source?
UTD Office of Development’s database on alumni giving.
2. What is the variable to be predicted?
Alumni giving. More specifically, whether an alumn is likely to give in FY2022.
4. Why traditional statistical method is better or not better?
Machine learning generally outperforms traditional statistical methods when it comes to prediction. This is because machine learning algorithms are able to improve automatically through experience based on the knowledge acquired from a dataset (generally referred to as the training set). On the contrary, traditional statistical models depend on manual programming where statisticians/programmers create programs (i.e. program improvement requires manually formulated and coded rules).