Final Year Project Suggestions - 2007
I am happy to discuss any project ideas in the areas of Text Data Mining, Machine Learning, AI, Information Retrieval.
Recent research into Textual Case-Base Reasoning (Delany & Bridge 2006) has shown that comparing documents using a similarity measure that is based on file compression (using the GZip algorithm) rather than the overlap in the words is very successful. This project would investigate whether using image compression techniques to compare images in a CBR system would be as successful.
The 9th European Conference on Case-based Reasoning (ECCBR 08) has issued a Call for Participation in the 1st Computer Cooking Contest. Details are available at http://www.wi2.uni-trier.de/eccbr08/index.php?task=ccc
This project would involve designing and implementing a system for entry to the competition.
Sentiment analysis aims to determine the attitude of the author of a document with respect to some topic. The challenge of this project is to identify whether it is possible to label textual documents according to whether they have a positive or negative attitude. For example, newspaper articles can be supportive or negative about some topic or film reviews can be positive or negative about a specific film. Specific applications of sentiment analysis include identifying racial material on the internet or using sentiment analysis to identify potential threats to a nation.
The project will involve building a system that can distinguish positive and negative attitudes in textual documents.
The International Journal of Meteorology wants to put together a website that will provide online access to the papers published in the journal. This is primarily a software engineering project that will involve the complete requirements analysis, design, and implementation of a comprehensive web application. The application should provide the following functionality registration mechanisms so that users can register with the journal and use the system;
In addition the project will require the student to investigate the following issues/constraints so as to provide as complete a solution to the Journal of Meteorology as possible:
The student who undertakes this project should consider using Ruby, the new reflexive, dynamic, fully object-oriented programming language and the framework Ruby on Rails to implement the web application.
Perudo or Liar's dice is a dice game. Each player starts with a cup and five dice of the same colour. A process of bluffing, bidding and, of course, luck, reduces the number of dice in play. The objective is to be the last player with any dice left. Rules are available here. In collaboration with the University of Paisley, I have an architecture available for playing the game, which includes players which use statistical techniques for determining their next move. A student last year designed a Perudo player which incorporated machine learning capabilities using Case-Base Reasoning which was successful at playing other players. This work has several ways of being extended, including using other AI techniques and extending the existing CBR functionality.
UNO is a card game for any number of players where each player much match the rank or suit of the previous card. There are a number of special cards with special effects such as reversing the direction of play, penalty cards and joker cards. The objective is to be the first player to lose all their cards. Rules are available here. The project objective is to implement the game of Uno and to evaluate whether "intelligent" selection of cards from a player's hand is better than random selection.