Data mining and Computational Statistics
Tentative list of possible areas of investigation for poster projects
Poster projects must be intended as an opportunity for student groups to investigate data mining/computational/statistical simulation problems and applications that interests them. The projects should apply data mining and computational statistics techniques learned during the course to real-world problems. Techniques other than those within the course syllabus can also be used, but we strongly recommend having a short discussion with the instructors before deciding to use the chosen technique/methodology. Data for these projects can be obtained from various internet sites, developed by students or delivered by the instructors (and listed in a separate document). R is a compulsory tool to be used within the projects. In what follow you will find a tentative list of possible topics which will be updated continuously in the next days.
FINANCE
• Bank customer credit scoring and profiling
• Stock market classification and forecasting
• Principal component regression of time series with many predictors
• Random variable generation for forecasting
• Credit card fraud detection
MARKETING
• Customer segmentation
• Customer retention
• Customer churn
• Customer care strategies
LINGUISTICS AND DOCUMENT CLASSIFICATION
• Web mining (including web usage, web structure and web content mining)
• Regression tree-based linguistics
SOCIAL NETWORKS AND INTERNET
• Wikipedia data
• Twitter data
PUBLIC OPINION MODELING
• Analysis of public opinion survey data
• Traditional and new media users' modelling
INFORMATICS
• Image recognition
• Intruder detection in TLC systems
• Public offering prices
PHYSICAL SCIENCES
• Weather data
MEDICINE AND BIOMEDICAL SCIENCES
• Decision making in medicine
• Effect analysis of new drugs and treatment