| title: | Data Mining Methods and Applications (DMMA) |
|---|---|
| reg no: | ETF5722 |
| project type: | Estonian Science Foundation research grant |
| subject: |
1.2. Applied Mathematics 2.9. System Engineering and Computer Technology |
| status: | accepted |
| institution: | University of Tartu |
| head of project: | Jaak Vilo |
| duration: | 01.01.2004 - 31.12.2006 |
| description: | In the Data Mining Methods and Applications (DMMA) project we will develop new Data Mining algorithms, techniques, and methods, mainly for the large-scale analysis of data from biomedical domain. These data are the primary sequences of DNA, RNA, and proteins, numerical data from high-throughput microarray gene expression measurements, scientific literature abstracts (text mining for extracting meaningful structure information from unstructured text), as well as several other types of experimental high-throughput molecular biology data. Additionally, we will explore data collected from medical records, lifestyle information, and genetic markers. The overall goal of the DMMA project is to develop methods suitable for large-scale analysis of such diverse data types and to satisfy various analysis needs. The main research questions that will be asked during each DMMA subproject are all those of a typical data mining project. We will start from studying the analysis needs of a particular application domain; combine and clean up data from different sources; decide about feature selection, knowledge representation formalisms, and evaluation criteria for the algorithms (MDL); explore the suitability and develop further the actual analysis algorithms (clustering methods, pattern discovery from the biosequences, and machine learning methods); and develop postprocessing methodologies and visualization techniques having end-users in mind. The proposed DMMA project is methods development project, the primary objective is novel algorithm development from the computer science viewpoint and data analysis techniques development from the data analysis viepoint. Applications enable to identify most urgent needs for the methods development. |
| project group | ||||
|---|---|---|---|---|
| no | name | institution | position | |
| 1. | Jaak Vilo | University of Tartu | Associate Professor | |