Data Mining

There are three main classes of possible topics for the research paper. These are elaborated below. Please remember that these are only suggested topics. If you have an idea for a paper which doesn't fit the three classes below, please contact me. I am very open to suggestions.

1. A paper on a particular DM technique

You should choose a data mining technique which interests you. There are many to choose from:

• Mining association rules
• Link analysis
• Memory-based reasoning
• Clustering/database segmentation
• Instance-based learning
• Logic programming
• Classification
• Estimation
• Prediction
• Decision trees
• Neural networks and self-organizing maps
• Genetic algorithms
• Deviation detection
• Rough sets
• Fuzzy sets
• Bayesian methods

(Note: the above list is not complete and the items are not necessarily independent.)

For the selected technique, you should give a basic theoretical explanation and consider its variants, usefulness and future directions.

2. A paper on the application of DM in a particular domain

You may choose a particular domain and review the application of data mining in that domain. Possibilities include:

• Customer relations
• Market analysis and management
• Operational structure analysis
• Fraud detection
• Loan approval
• Risk management
• Analysis of medical records
• The Human Genome project
• DM and the world wide web
• DM and multimedia

You should consider which techniques have been applied, giving a brief explanation of each technique. You should review the successes and/or failures of these techniques, and their advantages and disadvantages. These should be justified by reference to the theory of the techniques. Future directions in the application of data mining to the chosen domain should also be addressed.

3. A paper on a broad issue in DM

There are many possible issues. Examples include

• Validation and verification of DM results
• Visualization for DM
• Relationships between statistical analysis and DM techniques
• Is OLAP a DM technique?
• From Data Warehousing to DM
• Security and privacy for DM
• DM and ethics
• The problem of scale in DM
• Preprocessing for DM
• Metadata for DM