Optimization of Vehicle State Fatigue Feature Based on Ant Colony Algorithm
A modified ant colony optimization (IACO) algorithm is proposed to optimize the fatigue characteristics of vehicle states in response to road traffic safety issues caused by fatigue driving. The feature selection problem is transformed into a fully connected undirected graph and fisher scores and maximum information coefficients are in-troduced to improve search efficiency and reduce feature redundancy. Fatigue features are extracted from the vehi-cle state parameters when drivers are driving and the fatigue features are optimized using the IACO algorithm, and thus the optimal subset of fatigue features are obtained. The experimental result shows that the SVM classification accuracy of fatigue features optimized by the IACO algorithm is 85. 6％, and the KNN classification accuracy is 83.2％, both of which are higher than the classification results of other commonly used feature optimization algo-rithms. This indicates that the optimization performance of the IACO algorithm on fatigue features is higher than that of other commonly used feature optimization algorithms.
万方数据
维普
