Towards an intelligent tutoring system for database normalization: Design and evaluation of a decision support system
Database concepts are at the core of the graduate and undergraduate computing curriculum. Anecdotal evidence shows that understanding and appreciating the full breadth of this topic is usually difficult for many beginning students. This research develops a generalized decision support system model t...
Saved in:
| Main Author: | |
|---|---|
| Format: | Dissertation |
| Language: | English |
| Published: |
ProQuest Dissertations & Theses
01.01.2005
|
| Subjects: | |
| ISBN: | 0542398338, 9780542398339 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Database concepts are at the core of the graduate and undergraduate computing curriculum. Anecdotal evidence shows that understanding and appreciating the full breadth of this topic is usually difficult for many beginning students. This research develops a generalized decision support system model that would be used to provide supplemental pedagogical services for training database students. We are interested in stochastic problems where the environment of the problem-solving agent is uncertain. We focus on finding the closure of n attributes in an arbitrary database schema r(R) subjected to a given set of functional dependencies F. This well-known theory problem is relevant due to its impact in the area of efficient database design. Our proposed pedagogical database closure system relies on Bayesian modeling techniques and an object-oriented programming language for its implementation. The architecture of our system is centered on a fully-expanded "reasoning" tree representing all the choices the subject may take when constructing the path towards the solution of the problem. For each promising node in the tree, new nodes are connected to the ancestor as a result of using the each of the Armstrong axioms in an ordered way. The efficiency of our model is in the order of O(nm). The total efficiency of the model is exponential to the depth of the tree. We argue that our decision support system model can provide both a process- and product-centric solution space and the corresponding support for remedial actions in a wider sense than previously suggested by the literature. Learners can step through the system to learn the process to apply axioms to the current set of functional dependencies. Remediation could be offered at any point in the path. Alternately, because the system has solved all the possible paths to the final answer, the solution space can just as easily be evaluated. We could provide the ability to remediate the path to a solution if the learner has traveled "off-path" or did not provide a correct solution. |
|---|---|
| Bibliography: | SourceType-Dissertations & Theses-1 ObjectType-Dissertation/Thesis-1 content type line 12 |
| ISBN: | 0542398338 9780542398339 |