Advanced Fuzzy Clustering and Decision Tree Plug-Ins for DataEngineTM

Although a large variety of data analysis tools are available on the market today, none of them is perfect; they all have their strengths and weaknesses. In such a situation it is important that a user can enhance the capabilities of a data analysis tool by his or her own favourite methods in order to compensate for shortcomings of the shipped version. However, only few commercial products offer such a possibility. A rare exception is DataEngineTM, which is provided with a well-documented interface for user-defined function blocks (plug-ins). In this paper we describe three plug-ins we implemented for this well-known tool: An advanced fuzzy clustering plug-in that extends the fuzzy c-means algorithm (which is a built-in feature of DataEngineTM) by other, more flexible algorithms, a decision tree classifier plug-in that overcomes the serious drawback that DataEngineTM lacks a native module for this highly important technique, and finally a naive Bayes classifier plug-in that makes available an old and time-tested statistical classification method.