Main page

Registration/
Application

Timetable

Participants

Accomodation

Soft Methods and Tools in Data Analysis

Short description of the course:

Soft computing technologies are very important nowadays. Fuzzy sets, genetic algorithms, simulation of evolution processes or neural approaches appear everywhere in the scientific computation.
They constitute the basis of new application areas like data mining, costumer relationship management, simulation of dynamic systems that enable processing the economic, biological- , environmental systems etc.

Fuzzy systems

On the course we give a new sight to the fuzzy concept, which helps to understand the general idea of soft computing.

Datamining and queries

In datamining the most important thing is to understand what is behind the numerical values and categories. Fuzzy queries and sentence generator give an excellent solution for the problem.

Datamining and dependencies

To understand the data in a database is not other than discovering the relationships between the characteristics of the items. In fuzzy theory the operators play an important role. One of the most interesting ones is the Frank family. The measure property of this operator gives us the possibility to measure the dependencies between the categories. We show, how we can calculate it and use for data analysis.

Decision tree

Decision tree is very powerful tool to learn historical data or expert knowledge. ID3 works well on discrete categories. C4.5 handles numerical values on a very restrictive way. On the course we give a very general concept using the so called pliant concept and fuzziness measure.

Optimization and decision.

All kind of data analysis is for making good decision. Scoring is the most used decision support system. We answer the question how to establish such a system consistent with our experience.

New ideas

At the end of the course we summarize the solutions and outline of some new results.

Preparatory reading

I suggest first of all my articles on DeMorgan Class of fuzzy operators, aggregation operators and membership functions. Some of them can be downloading from my webpage Dombi Jozsef

Articles by Jozsef Dombi:
1. Membership function as an Evaluation (ps-format) ,
2. A General Class of Fuzzy operators, the DeMorgan Class of Fuzzy Operators and Fuzziness Measures Induced by Fuzzy Operators (ps-format) ,

I can recommend some other books as well:

J. Fodor, M. Roubens, Fuzzy preference modeling and multicriteria decision support.
D.Dubois, H. Prade: Fundamentals of fuzzy sets
E.P.Klement, R.Mesiar, E.Pap: Triangular norms
J.R. Quinlan: C4.5 Programs for machine learning
Carlos A. Bana e Costa: Readings in Multiple criteria decision making.

Additional material for reading

Learning multicriteria classification models from examples: Decision rules in continuous space (pdf-format), (ps-format).


The student will use on supervised exercises Maple and Excel (maybe also some nonlinear optimization tool).