A comparative study of machine learning models for fundraising success

No Thumbnail Available
Nabar, Rajas
Issue Date
MSc in Data Analytics
Dublin Business School
Items in eSource are protected by copyright. Previously published items are made available in accordance with the copyright policy of the publisher/copyright holder.
Non-Profit organizations play a vital role in mitigating the plights of society by providing nec-essary services. However, to accomplish their goals they need an ongoing infusion of funds. Prospective donor identification and donor amount upgradation are two major challenges faced by these organizations. The research overcomes these challenges by a comparative study of seven machine learning models which are Gradient Boosting, XGB, KNN, SVM, Naïve Bayes, Logistic Regression, and Decision Tree. As per the research, XGB performs the best overcom-ing both the challenges by achieving an accuracy of 89% and F1-Score of 88% for donor pre-diction and an accuracy of 95% and F1-score of 91% for donor amount upgradation. The re-search also helps to deal with a class imbalance which is a common issue with classification problems by using the SMOTE technique. Further, an interactive web application is built using R-Shiny to help the fundraisers in their mission.