An Applied Method for Knowledge Acquisition

摘要

This paper presents an applied approach for knowledge acquisition, which is applied to an optimal control of a part transmission system. Firstly, simulation technique is adopted in this paper for collecting original data. Secondly data preprocessing is performed, based on a statistical method, on the original data to remove condition attributes, which are redundant and have little importance to the decision attributes, and the scale of the data set can be reduced. Thirdly, a modified genetic algorithm combined with the nearest neighbor distance method is used for clustering the preprocessed data set. As a result, simple, credible knowledge rules can be obtained. Experimental simulation results show that these knowledge rules used for optimal control of the system under study are more effective.