Title

Analysis of CWU foundation data: pre-processing and data mining

Presenter Information

Ying Wang

Document Type

Oral Presentation

Location

SURC 137B

Start Date

17-5-2012

End Date

17-5-2012

Abstract

The CWU foundation is an organization which works with donors, alumni, and friends to raise private funds to support CWU students, faculty, and programs. It has accumulated over 20 year's data up to year 2000 about their members and donations. The CWU foundation hopes to find predictive patterns from those data and to use them as guidelines to raise more funds. In this project, various data preparation methods and mining algorithms were used to accomplish the task. Based on the independent quality of the data, naïve Bayes classifier was chosen because of its simplicity and performance. Confusion matrix was used to evaluate the performance of the classifier. Half of the data was used to build a model of probability distribution; the other half was used to test the model. The results show that two types of members are more likely to donate to CWU. One is married members with double income; the other is members who work in educational areas. The future step is to gather more data after year 2000 and find more behavioral patterns of donors with time series analysis.

Faculty Mentor(s)

Boris Kovalerchuk

Additional Mentoring Department

Computer Science

This document is currently not available here.

Share

COinS
 
May 17th, 9:30 AM May 17th, 9:50 AM

Analysis of CWU foundation data: pre-processing and data mining

SURC 137B

The CWU foundation is an organization which works with donors, alumni, and friends to raise private funds to support CWU students, faculty, and programs. It has accumulated over 20 year's data up to year 2000 about their members and donations. The CWU foundation hopes to find predictive patterns from those data and to use them as guidelines to raise more funds. In this project, various data preparation methods and mining algorithms were used to accomplish the task. Based on the independent quality of the data, naïve Bayes classifier was chosen because of its simplicity and performance. Confusion matrix was used to evaluate the performance of the classifier. Half of the data was used to build a model of probability distribution; the other half was used to test the model. The results show that two types of members are more likely to donate to CWU. One is married members with double income; the other is members who work in educational areas. The future step is to gather more data after year 2000 and find more behavioral patterns of donors with time series analysis.