Multiple attribute dynamic fuzzy decision tree approach for voltage collapse evaluation

Voltage collapse is a complex phenomenon which has a variety of contributing factors. Past efforts have been given in analysing this phenomenon. As a result, various methods of analysis have been devised. Some methods are considered to be complex, slow but accurate and some methods are considered to simple, fast but inaccurate. With the emergence of machine learning techniques, a data mining method can also be used as an alternative diagnostic tool. This method is known as fuzzy decision tree. This paper will outline improvements made to an existing fuzzy decision tree method by adding more contributing attributes for partitioning, creating a hybrid fuzzy decision tree. Comparison and tests are made using an IEEE 300 bus system.