Research Projects
The research activities in the competence center are organized in concrete projects. The projects investigate specific problems in the area of medical data warehousing and analysis which emerge from the realization of a clinical data warehouse in the hospital Meran.
Pattern-Based Temporal Selection in Clinical Data Warehouses. This project aims to develop a new temporal selection operator, in which the selection of tuples is triggered by the occurrence of events which can be described as a complex temporal pattern. This kind of tuple selection facilitates the analysis of huge amounts of chemotherapy data, which are characterized by long histories of events such as medications and examinations.
MOBAS. The overall objective of the MOBAS project is to extend the data analysis and decision support functionality of the Day Hospital for Onco-Hematology in two directions: a) mobile services for context-aware information management and b) decision support and patient medical data analysis and mining. Mobile services will enable the clinician to send context-dependent messages to the patients, regarding: their next planned activities, information about side-effects of the therapy, or inviting them to fill a questionnaire. Personalization techniques will be used to detect the most suitable situation for the patient to receive such messages by analyzing log data of previous interactions.
k-NN Search for Chemotherapy Histories. The project aims to investigate k-NN queries for patietns with long chemotherapy treatment histories. Given a specific patient P, find the k patients P1,...,Pn with a chemotherapy history which is most similar to the history of P. This task requires first to define a similarity measure for patient histories and second to provide efficient evaluation algorithms.
In order to assist in developing search strategies and similarity measures for treatment histories, Jay Anderson provided a variety of visualizations of courses-of-treatment in order to have a visual comparison of treatments claimed to be similar. More information about the visualization methods is here.
Detecting Hidden Relationships in Clinical Lab Data. The aim of this project to analyse clinical lab data to detect hidden relationships between the various parameters (disease, test results, age, etc.). We will apply various data analysis/mining and data visualization tools including the 3VDM system.
