Postgraduate Project Ideas:

• Automatic Identification of Data Objects/Classes from Requirement Specifications
  • Database design is usually performed manually by expert database designers/developers.
  • The process could be greatly simplified and make efficient if we identify objects/classes of interest to the domain model from requirements specifications, which are usually written in plain or structured English.
  • There are different techniques used for the task e.g., noun and noun phrases.
  • The project involves software engineering, object-oriented analysis and design, and database modelling and design.
• Database Design for Object-Relational Systems
  • There are quite a few methods/approaches for database design for relational databases. However, not much is available in terms of theory and practice, let alone a method, for database design for object-relational databases.
  • The project seeks to investigate ways in which an object-oriented conceptual database model could be turned into logical/physical object-relational design for implementation using an object-relational database management system.
• Main Memory Databases
  • There is still a substantial gap between searching in and data retrieval from main memory and disk.
  • Data processing can only occur when data is fetched from disk and stored into main memory. Now-a-days, main memory is becoming larger and larger e.g., 1GB is normal for a desktop PC.
  • An average size database is not more than a couple of GBs.
  • This project will try to investigate ways in which the entire database could be copied to the main memory for query processing and updates.
  • The emphasis would be on whether this will improve database performance.
• Eager/Proactive  Optimisation
  • Traditionally, a query optimiser of a database system suggests ways in which certain queries (e.g., written in SQL) could be optimised such that less resources (memory, CPU, disk input/output) are spent and to speed up response time. However, its capabilities to do so depends entirely on how the database in question is structure and what access paths (e.g., indexes, partitions, clustering, materialized views) are available.
  • Eager or proactive optimisation is about analysing existing database structures and ad hoc queries and suggesting ways in which the underlying database structures could be altered, optimised, improved in order to achieve optimal performance.
  • This could be carried out in different modes e.g., suggestive, transparently and automatic or negotiative /consultative.
• Personal Digital Image Cataloguing
• Over time we take, collect and receive hundreds and thousands of digital images.
• Looking for images of interest in a large collection is time consuming and frustrating.
• There is a need for a tool that could make storage and retrieval of images easy and efficient.
• The project requires investigation into ways in which images could be stored in a database such that they could be later retrieved using simple and easy to recall searching criteria.
• You may decide to look into different ways of classifying images (e.g., event, location, people, etc). Moreover, you may decide to investigate image retrieval by content or metadata or keywords.

Related Links: If you want to look into what goes on in the feild of databases, visit some of the following links:

http://www.informatik.uni-trier.de/~ley/db/conf/sigmod/index.html

http://www.informatik.uni-trier.de/~ley/db/conf/vldb/index.html

http://www.informatik.uni-trier.de/~ley/db/conf/edbt/index.html

http://www.informatik.uni-trier.de/~ley/db/conf/bncod/index.html

http://www.informatik.uni-trier.de/~ley/db/index.html