The main activity of Biostatistics Group is elaboration of statistical analysis (according with the European Union standards) for medical research and experiments based on clinical databases.
Research problems are focused on four main issues concerning the application and developing of statistical methods in different branches of science (e.g. medicine, genetics, medicinal chemistry):
- Statistical techniques of filling in missing data.
The main aim of the project is to study the influence of filling in missing data with the help of innovative statistical methods on the results of the research. The efficacy of these methods will be verified based on the model used in the identification of speaker. The frequency of four the lowest formants describing the acoustic structure of vowels will be used in the construction of the model. The analyzed dataset contains trial recordings of five speakers. The demographic features of the speakers are similar (man in the age 25-28) living in the same place.
- The study of changes in the electronic structure of small (non-protein) molecular systems due to their interactions with the proteins.
The project focuses on searching for chemical compounds or their groups (e.g. carboxylic acids, monosaccharides, vitamins), which electronic structure changes significantly after formation of a complex with a protein. The next aspect connected with this problem lies in the electronic properties (accessible from quantum-chemical calculations like total energy of the system, atomic partial charges, electrostatic potentials) of this compounds, which changes as well in the reference to the values obtained for their isolated state. The final results will require application of advanced multidimensional statistics methods.
- Elaboration and implementation of the methods applied in the frame of Data Mining for medical science.
The main goal of the Project is implementation and modification of selected Data Mining techniques to medical diagnostics. The artificial intelligence and statistics tools are planned to used in discovering valuable information hidden in medical databases. Medical data analysis with the application of Data Mining or artificial neural networks can be helpful yielding small number of rules with high diagnostic values. Data Mining applicability facilitates studies of new genes function and designing new drugs enabling proper treatment.
- Medical data analysis of clinical data sets.
Application of multidimensional statistics and neural networks to selected classification and prediction problems.