Research on Multi-Sensor Feature Fusion Based on Improved CBPSO-SVM in WSN

Provided by: Binary Information Press
Topic: Networking
Format: PDF
In pattern classification of Wireless Sensor Network (WSN), multi-senor feature fusion can improved the classification performance, but there are lots of irrelevant and redundant features which cause considerable computation even restrict the performance. To solve this problem, a wrapper feature selection approach combining Constrain-Based Particle Swarm Optimization (CBPSO) and Support Vector Machine (SVM) for multi-sensor feature fusion is proposed. Improved CBPSO algorithm is utilizing to select an optimum subset containing more effective features for classification, while SVM with k-fold cross-validation is used to evaluate the classification accuracy and efficiency of feature selection.

Find By Topic