Optimization of PBFT Algorithm Based on Improved C4.5

Aiming at the problems of PBFT algorithm consortium blockchain, such as high communication overhead, low consensus efficiency, and random selection leader nodes, an optimized is proposed. Firstly, improves C4.5 introduces weighted average information gain to overcome mutual influence between conditional attributes improve classification accuracy. Then classify nodes with improved C4.5, select ones a trust level form main group. Finally, integral voting mechanism introduced determine node. Experimental results show that compared traditional algorithm, times are reduced greatly, which effectively alleviates problem number in increases traffic volume too large, significantly reduces probability node doing evil efficiency.