An Improvement Forward Floating Search Algorithm for Feature Selection in Power System Transient Stability Classification

This paper proposed feature selection algorithm in power system transient stability classification. Feature selection is a very important stage aimed at reducing a number of features, retaining distinctive features to reduce memory, increasing identification accuracy. The proven effectiveness of Improvement Forward Floating Search algorithm (IFFS) was compared with Sequential Forward Selection (SFS), Sequential Forward Floating Selection algorithm (SFFS) and Relief algorithm through analysis of results in power system transient stability classification using K-Nearest Neighbour (K-NN) on IEEE 30-bus standard power network. The analytical results showed that the IFFS achieved effective reduction features and recognition accuracy higher than other methods.