Predicting Donors - KDD Cup 1998 Data

Problem Statement - Predict if the previous donors of a national veteran organization are going to donate again.

Data Source

Installation

Download the data

• Clone this repo to your computer.
• Get into the folder using cd PredictingDonors.
• Download the data file 'cup98lrn.zip' from the data source mentioned above
• Extract the text file from the zip folder and open it in excel.
• Save the file as 'cup98lrn.csv' (comma delimeted csv file format).
• Remove all the zip files by running rm *.zip.

Install the requirements

• Make sure you use Python 3.
• Install all the required libraries. (or choose an easy life and use Anaconda :D )

Usage

• Run python Donors.py.
  • This will run 4 predictive models, and print performance of each model measured in 4 metrics
    • Recall
    • F1
    • Precision
    • Accuracy

Extending this

If you want to extend this work, here are a few places to start:

• Run regression on the scored dataset generated from 'Donors.py' and estimate return from direct mailing to maximise donation profits.

Current Result