Study on information fusion algorithm and application based on improved SVM

Authors presented the information fusion algorithm based on improved SVM, namely, decision tree - support vector machine algorithm (Decision Tree Method-Support Vector Mechines, DTM-SVM). The algorithm overcame the limitations of the conventional SVM classification which applied only to two-classification problem by a “one to many” pattern, solved multi-classification problem and met a wider range of application requirements. Finally, based on the establishment of a freeway traffic state identification evaluation system, the DTM-SVM model was applied to solve the freeway traffic state recognition. Results show that: the algorithm can identify in a shorter time to reach higher recognition accuracy.