Security Evaluation in power operation is presented as a pattern recognition problem . An important element in pattern recognition is related to the selection of effective feature from a larger set of feature measurements .thus reducing the complexity and size of the decision function.
The problem of feature extraction has been treated in this paper using clustering technique of the Karhunen-loeve(K-L) expansion type. A set of optimum features has been derived through a clustering transformation (K-L) technique, these feature vectors are used to form the orthogonal transformation matrix to derive the new image patterns.
As the feature extraction operation is established .the study of dynamic security is completed by identifying the decision. And the percent misclassification due to different techniques.
A case study was under taken using the proposed technique and, two other traditional techniques, namely the Single Ranking (SR), and Repetitive Ranking (RR) techniques.