Sanchez, Carlos
(2016)
An Analytics Based Architecture and Methodology for Collaborative Timetabling in Higher Education.
Doctoral Dissertation, University of Pittsburgh.
(Unpublished)
Abstract
Class scheduling in higher education, also known as “timetabling”, is a complex process that involves many people across an institution for several months every year, and literature on the topic has been rapidly evolving over the last 15 years. We propose architecture and methodology to enable the implementation of systems that can help users gain insight on non-trivial existing and emerging enrollment patterns that need to be considered for planning purposes, and to facilitate collaborative timetabling activities. University of Pittsburgh data on undergraduate enrollments during six recent fall terms is used to illustrate the proposed ideas. Core components are specified by: First, modeling the problem using Association Rule Analysis where the sets of courses that individual students take in an academic term are treated as transactions. This renders combinations of courses called itemsets. A new backtracking algorithm called MASAI is proposed to determine the maximum number of seats available per itemset. This corresponds to the identification of itemsets of interest as in the case at hand course itemsets with no seats available are primary targets. MASAI is a novel approach to the identification of itemsets of interest that uses information that is not available in transactional data to determine the maximum number of seats possible in each itemset. Second, in order to facilitate deeper analyses that consider the relationships between course itemsets, the problem is modeled as a multi-mode graph that incorporates information obtained with the Association Rule Analysis and MASAI. A Generalized Clique Percolation Method (GCPM) is proposed to enable the identification of overlapping and hierarchical communities in graphs/networks. GCPM is used to identify communities in the multi-mode graph, enabling the discovery of non-trivial enrollment patterns, and the identification of scheduling practices that limit the enrollment options for students. Third, the elements that would form the core of a socially translucent environment that is based on the previous components are discussed. This collaborative environment is intended to provide scheduling authorities with access to shared information on enrollment patterns and how decisions on scheduling of courses in their departments impact the overall institution’s schedule and the enrollment options for students.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
| Item Type: |
University of Pittsburgh ETD
|
| Status: |
Unpublished |
| Creators/Authors: |
|
| ETD Committee: |
|
| Date: |
13 January 2016 |
| Date Type: |
Publication |
| Defense Date: |
2 December 2015 |
| Approval Date: |
13 January 2016 |
| Submission Date: |
6 December 2015 |
| Access Restriction: |
No restriction; Release the ETD for access worldwide immediately. |
| Number of Pages: |
184 |
| Institution: |
University of Pittsburgh |
| Schools and Programs: |
School of Information Sciences > Information Science |
| Degree: |
PhD - Doctor of Philosophy |
| Thesis Type: |
Doctoral Dissertation |
| Refereed: |
Yes |
| Uncontrolled Keywords: |
Algorithm, Association Rules Analysis, Clique Percolation Method, Collaborative Timetabling, Community Analysis, Community Structure, Generalized Clique Percolation Method, Graphs, Group Decision Support Systems, Hierarchical Communities, Higher Education, Market Basket Analysis, Negative Association Rules, NEO4J, Networks, Overlapping Communities, Rules of Interest, Scheduling, Social Translucence, Timetabling |
| Date Deposited: |
13 Jan 2016 16:15 |
| Last Modified: |
15 Nov 2016 14:31 |
| URI: |
http://d-scholarship.pitt.edu/id/eprint/26609 |
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[img]](https://d-scholarship.pitt.edu/style/images/fileicons/application_pdf.png)
![[feed]](/images/feed-icon-32x32.png){kind=icon}