Frequently asked questions
- What is the Decision Master?
- What can the Decision Master be used for?
- For which applications has the Decision Master been used so far?
- On which Data-mining Techniques is the Decision Master based?
- Of which Options does the Decision Master Dispose?
- Which Possibilities of Data Processing does the Decision Master have?
- How are the Results Evaluated?
- Where can I find Theoretical Information on these Techniques?
- What Distinguishes the Decision Master from other Data Mining Software?
- Can the Decision Master also be operated by Persons who are no Computer-scientists?
- In which Format must the Data be that are to be Analysed?
- Is the Decision Master Developed Further?
Decision Master is a software tool for the automatic extraction of non-trivial contexts in large data collections for prediction, diagnosis and recognition of new knowledge. The learnt/acquired knowledge is represented in the form of decision trees or rules. After learning the knowledge can be evaluated with secured statistical techniques. In the case of sufficient quality of the knowledge it can be used for diagnosis by Decision Master. home
Decision Master can be used for learning decision knowledge in different fields. The prerequisite is that a database is available with information on the specific field. This database is used by Decision Master to extract generalizing information from the database. home
The spectre of applications for which Decision Master has been used is very wide. Decision Master has been applied in medicine, non-destructive testing, pharmacy, marketing, biotechnology and other fields. home
Decision Master is based on induction techniques. Decision trees are learnt. To this end a multitude of different induction techniques are put at disposal. In-house developed decision-tree techniques as e.g. HD 2.0, EDD 2.0 or VED 2.0 as well as standard techniques like e.g. C4.5 or ID3 were realized. Binary as well as polyvalent decision trees can be learnt. home
The technique for learning the decision tree can be compiled freely. There are options for
- the way of attribute selection during the learning of the decision tree
- the way of discretization of numerical attributes
- the technique for pruning and
- the technique for evaluating the decision tree. home
Decision Master is able to do tear-out detection and analyse the range of values. home
The results can be evaluated by test-and-train or cross validation. home
Theoretical information on the techniques can be found in the book P. Perner, Data Mining on Multimedia Data, Springer Verlag, LNCS 2558, ISBN: 3-540-00317-7, 2003. home
Decision Master has a user-friendly surface. No special computer knowledge is required. for the employment of Decision Master. The multitude of possible techniques is particularly suited for experiments. home
Yes. The program is menu-driven and supports all the steps that are necessary to carry out a Data Mining experiment. home
The data must be available in MS Excel format (*.xls), in DBase format (*.dbf) or as comma separated textfile. home
Yes. Decision Master is continually attended by our team/group and developed further. home