Abstract: Ensemble learning is a machine learning paradigm where multiple models arc strategically generated and combined to obtain better predictive performance than a single learning method. It was proven that ensemble learning is feasible and tends to yield better results. Uncertainty reasoning is one of the important directions in artificial intelligence. Various uncertainty reasoning methods have been developed and all have their own advantages and disadvantages. Motivated by ensemble learning, an ensemble method of uncertainty reasoning was proposed The main idea of the new method is in accordance with the basic framework of ensemble learning, where multiple uncertainty reasoning methods is used in time and the result of various reasoning methods is integrated by some rules as the final result. Finally, theoretical analysis and experimental tests show that the ensemble uncertainty reasoning method is effective and feasible.

Key words: Uncertainty reasoning, Mixed reasoning, Ensemble learning