A Case Study for the Application of Data and Process Mining in Intervention Program Assessment and Improvement

Elnaz Douzali; Houshang Darabi

Download Paper | Permalink

Conference: 2016 ASEE Annual Conference & Exposition
Location: New Orleans, Louisiana
Publication Date: June 26, 2016
Start Date: June 26, 2016
End Date: June 29, 2016
ISBN: 978-0-692-68565-5
ISSN: 2153-5965
Conference Session: Student Success I: Interventions and Programs
Tagged Division: Educational Research and Methods
Page Count: 15
DOI: 10.18260/p.26267
Permanent URL: https://peer.asee.org/26267
Download Count: 1243

Paper Authors

biography

Elnaz Douzali University of Illinois, Chicago

visit author page

Elnaz Douzali is a senior undergraduate researcher at the University of Illinois at Chicago. She's a part of the Mechanical and Industrial Engineering Department and will receive her Bachelors of Science in Industrial Engineering in May 2016. Since 2015 Elnaz has participated in multiple projects in Educational Data Mining. Her research interests include Educational Data Mining, Process Mining, and Healthcare. Elnaz will begin her Masters of Science in Industrial Engineering at the University of Illinois at Chicago in the fall of 2016.

visit author page

biography

Houshang Darabi University of Illinois, Chicago orcid.org/0000-0001-7881-6542

visit author page

Dr. Houshang Darabi is an Associate Professor of Industrial and Systems Engineering in the Department of Mechanical and Industrial Engineering (MIE) at the University of Illinois at Chicago (UIC). Dr. Darabi has been the Director of Undergraduate Studies in the Department of MIE since 2007. He has also served on the College of Engineering (COE) Educational Policy Committee since 2007. He is currently the Director of Analytics and Capacity Planning at the COE.
Dr. Darabi is the recipient of multiple teaching and advising awards including the COE Excellence in Teaching Award (2008, 2014), UIC Teaching Recognitions Award (2011), and the COE Best Advisor Award (2009, 2010, 2013). Dr. Darabi has been the Technical Chair for the UIC Annual Engineering Expo for the past 5 years. The Annual Engineering Expo is a COE’s flagship event where all senior students showcase their Design projects and products. More than 600 participants from public, industry and academia attend this event annually.
Dr. Darabi is an ABET IDEAL Scholar and has led the MIE Department ABET team in two successful accreditations (2008 and 2014) of Mechanical Engineering and Industrial Engineering programs. Dr. Darabi has been the lead developer of several educational software systems as well as the author of multiple educational reports and papers. Some of these products/reports have already been launched/completed and are now in use. Others are in their development stages. Dr. Darabi’s research group uses Big Data, process mining, data mining, Operations Research, high performance computing, and visualization techniques to achieve its research and educational goals.

visit author page

Download Paper | Permalink

Abstract

In this paper, we demonstrate the details of a case study in which we used data mining and process mining methods to assess the success of an intervention method on the student population within ABC University. In addition to assessing the intervention method, we also suggest a set of decision making rules that if supplemented with the current intervention method can potentially improve the assessment outcomes.

Within ABC University, a proportion of undergraduate students who satisfy a minimum Cumulative Grade Point Average (CGPA) of 3.4 (out of 4.0) are placed in a special intervention program called DEF. DEF participants are offered general education courses with smaller class sizes, are assigned additional counselors, and are provided more scholarship opportunities. All DEF participants are monitored after each semester and if a DEF participant’s CGPA decreases to lower than 3.4, the participant is dropped from the intervention program.

The success of the DEF intervention program is measured by the percentage of its participants who keep their CGPA above 3.4 (called CGPA criterion) as well as the percentage of its participants who graduate from ABC university within 6 years of beginning (called graduation criterion).

Our study included the analysis of more than 17,000 students of ABC University who entered the university as freshmen between 2005 to 2009. Out of these 17,000 students, more than 2,500 attended the DEF intervention program for at least one semester out of their first four semesters. The remaining students were not exposed to the intervention program. We created a trace (a set of events) for each student by tracking their CGPA in their first 4 semesters. In each semester, if a student’s CGPA was lower than 3.4, we recorded an N (non-satisfactory) event for that student. A CGPA of 3.4 or higher generated an S (satisfactory) event. As the output (final) event for each student’s trace, we recorded a D (showing that the student dropped from the university) or a G showing that the student graduated. The D and G events did not necessarily occur in the first 4 semesters for each student. The D and G events were available for all students as our data set included only the students who entered ABC University as a freshman between 2005 to 2009.

We used visualization, data mining, and process mining tools to analyze the success of the intervention program. Our analysis revealed that the DEF program did not have a significant impact on the success of its participant who attended the program in their first semester. However, for groups of students with certain trace structures, the DEF program was very effective. For these groups, both the CGPA and the graduation criteria were significantly better compare to the students who were never exposed to the DEF program. Based on the trace structures that were significantly effective, we developed a set of rules to help the DEF administrators to decide how to admit or dismiss students from this program to improve the intervention success and effectiveness.

Citation
Format

Douzali, E., & Darabi, H. (2016, June), A Case Study for the Application of Data and Process Mining in Intervention Program Assessment and Improvement Paper presented at 2016 ASEE Annual Conference & Exposition, New Orleans, Louisiana. 10.18260/p.26267

TY - CPAPER
AB - In this paper, we demonstrate the details of a case study in which we used data mining and process mining methods to assess the success of an intervention method on the student population within ABC University. In addition to assessing the intervention method, we also suggest a set of decision making rules that if supplemented with the current intervention method can potentially improve the assessment outcomes.

Within ABC University, a proportion of undergraduate students who satisfy a minimum Cumulative Grade Point Average (CGPA) of 3.4 (out of 4.0) are placed in a special intervention program called DEF. DEF participants are offered general education courses with smaller class sizes, are assigned additional counselors, and are provided more scholarship opportunities. All DEF participants are monitored after each semester and if a DEF participant’s CGPA decreases to lower than 3.4, the participant is dropped from the intervention program.

The success of the DEF intervention program is measured by the percentage of its participants who keep their CGPA above 3.4 (called CGPA criterion) as well as the percentage of its participants who graduate from ABC university within 6 years of beginning (called graduation criterion).

Our study included the analysis of more than 17,000 students of ABC University who entered the university as freshmen between 2005 to 2009. Out of these 17,000 students, more than 2,500 attended the DEF intervention program for at least one semester out of their first four semesters. The remaining students were not exposed to the intervention program. We created a trace (a set of events) for each student by tracking their CGPA in their first 4 semesters. In each semester, if a student’s CGPA was lower than 3.4, we recorded an N (non-satisfactory) event for that student. A CGPA of 3.4 or higher generated an S (satisfactory) event. As the output (final) event for each student’s trace, we recorded a D (showing that the student dropped from the university) or a G showing that the student graduated. The D and G events did not necessarily occur in the first 4 semesters for each student. The D and G events were available for all students as our data set included only the students who entered ABC University as a freshman between 2005 to 2009.

We used visualization, data mining, and process mining tools to analyze the success of the intervention program. Our analysis revealed that the DEF program did not have a significant impact on the success of its participant who attended the program in their first semester. However, for groups of students with certain trace structures, the DEF program was very effective. For these groups, both the CGPA and the graduation criteria were significantly better compare to the students who were never exposed to the DEF program. Based on the trace structures that were significantly effective, we developed a set of rules to help the DEF administrators to decide how to admit or dismiss students from this program to improve the intervention success and effectiveness.

AU - Elnaz Douzali
AU - Houshang Darabi
CY - New Orleans, Louisiana
DA - 2016/06/26
PB - ASEE Conferences
TI - A Case Study for the Application of Data and Process Mining in Intervention Program Assessment and Improvement
UR - https://peer.asee.org/26267
DO - 10.18260/p.26267
ER -