Topics

In all human activities, automatic data collection pushes towards the development of tools that are able to handle and analyze data in a computer-supported fashion. In the majority of the application areas, this task cannot be accomplished without using the available knowledge on the domain or on the data analysis process. This need becomes essential in biomedical applications, since medical decision making needs to be supported by arguments based on basic medical and pharmacological knowledge. In this working group we will devote our study to computational methods for data analysis aimed to narrow the gap between data gathering and data comprehension, as well as their applications in medicine, health care, biology and pharmacology. Methods for analyzing data by integrating the available knowledge on the domain (Intelligent Data Analysis) and for extracting knowledge from large data-bases (Data Mining) will be both investigated. Therefore, the topics of the WG will include, but will not be limited to, effective machine learning and data mining tools, clustering, data visualization, case-based reasoning, interpretation of time-ordered data (derivation and revision of temporal trends and other forms of temporal data abstraction), outcomes research, construction of prognostic models to support medical decision making, discovery of new drug compounds, predicting drug activity, analysis of large biomedical data-bases such to assist in domains such as protein structure prediction and gene function prediction. Emphasis will also be given to solving of problems which result from automated data collection in modern hospitals, such as analysis of computer-based patient records, data warehousing tools, intelligent alarming, effective and efficient monitoring, and so on.

The Intelligent Data Analysis and Data Mining Working Group activity will be devoted to i) the computational methods for data analysis in medicine and pharmacology that are able to exploit the additional expert knowledge of the problem domain (Intelligent Data Analysis) and ii) the computational methods for data analysis able to extract information from potentially unstructured large data sets (Data Mining). Effective machine learning tools nowadays provide means to derive understandable diagnostic and prognostic rules; clustering, instance-based learning methods, like case-based reasoning, may provide crucial help to physicians in their decision making process; the interpretation of time-ordered data through the derivation and revision of temporal trends and other forms of temporal data abstraction provides a powerful instrument for situation-detection and prognosis; data visualization is more and more an essential part of the overall process of knowledge discovery in databases; data mining can extract useful relationships from large data-bases and data-warehouses which may point out to a potentially new and useful knowledge that was hidden in the data. Finally, Bayesian Networks and Fuzzy Systems represent well-known data analysis and reasoning tools able to explicitly deal with prior knowledge in uncertain domains. Special emphasis will be given to systems that aim at integrating the above mentioned methodologies to promote the construction of effective decision models to support medical decision making, discovery of new drug compounds, pharmacodynamical modeling, prediction of drug activity, protein structure prediction, analysis of gene expression data, and so on. Attention will also be given to solve problems which result from automated data collection in modern hospitals, such as analysis of computer-based patient records (CPR), data warehousing tools, outcomes analysis, intelligent alarming, effective and efficient monitoring, and so on. In particular, we will try to stress the following scientific issues:

• What are the application classes that motivate the usage of certain methods;
• What is the potential applicability (and generalizability) of proposed solutions;
• What is the level of integration with other methods and tools to achieve real working systems;
• What kind of knowledge is needed, used and/or extracted by the IDA and DM methods;
• What is the role of prior knowledge in data analysis;
• How should the available knowledge be represented.