Semi-supervised Learning Algorithm Based on Simplified Association Rules Combining with k-mean and Its Application in Land Evaluation

In order to construct intelligible and effective land evaluation classifier, a semi-supervised learning algorithm constructed by utilizing simplified association rules combining with k-mean clustering algorithm is proposed in this paper. To reduce the complexity of the land evaluation models and improve the efficiency and intelligibility of association rules further, an algorithm to eliminate redundant rules for obtaining the simplified association rules is presented. Experimental results of Guangdong Province land resource demonstrate that, by only using 500 training samples chosen randomly, 89.5143% correct area rate of land evaluation could be obtained by the semi-supervised learning algorithm. It provides a higher precision with the accuracy improved by 14.3484%, comparing with the results of the method k-mean and 7.1159% comparing with the results of the method support vector machine in the same condition.